Exam 2 Flashcards
Phonemes occur at the ____
Linguistic level
Sound is a _____
Pressure wave
String Fallacy
Speech sounds are separable and sequential
Speech segmentation is hard because:
- Phonemes overlap
- Phonemes affected by adjacent elements
Voice onset time
Time between consonant release and vocal chord vibration
Continuous perception
Notice every change in perception
Categorical Perception
Perception doesn’t always change just because physical properties do
- Between category is more effective
- Neonates have it
- Innate and domain general
Motor Theory
Acoustic of speech perception continuous, actions of speech perception categorical
-Only infants talking have categorical perception
Conditioned sucking
Infants react at a VOT of 20 for between category, and a VOT of 60 for within category
McGurk Effect
Phoneme categorization can be affected by visual info
-Infants dishabituate and seem interested in perceiving /da/ when audio is /da/
Hindi has 2 stops: ___ and ____
Dental and retroflex (English is between the two)
After 6/8 months, _____
infants-adults do not perceive differences in non-native language
Maintenance/loss view
Non-native boundaries disappear, native language becomes the spotlight, BUT children can learn new languages without accents
Functional Reorganization
Native language phonemes built from universal perceptual categories, helps to learn words
- Innate
- Phonological at 10 months
- Lexical at 14 months
- Syntactic at 18 months
Toddlers have difficulty learning words that change by ___ feature
One
What info creates representations? (3 hypotheses)
- Represent categories heard most often (frequency)
- Learning through presence of contrast (contrast)
3.
Learning by contrast
Bimodal= sensitivity between tokens 3 and 6 (two mountains) Monomodal= sensitivity between tokens 1 and 8 (one mountain)
Lexical distinctions—>___
Non-lexical distinctions—->____
Bimodal distribution
Unimodal distribution
Speech production stage 1
Reflexive vocalization
- birth to 8 weeks
- automatic responses to environment
- hunger cry, pain cry, temper cry
Speech production stage 2
Cooing
- 2-4 months
- intentional vowel-like sounds
- /e/ and /u/
Speech production stage 3
Isolated vowel-like sounds
- 4-7months
- pre-canonical babbling
- vocal experimentation
- vocal tract changes- soft palate shifts up, tongue separates from jaw
Speech production stage 4
Canonical babbling
- 7 months on
- reduplicated babbling
- true words
- variegated babbling 12 months on
- jargon babbling
Why do infants babble?
Motor development, language development
Babbling sounds
12 most frequent consonants make up 95%- stops, glides, nasals
Jaw oscillation theory
Alternating between open and closed vocal tract due to single action
-Vowels= resting tongue
Jaw oscillation consonants and vocal tract
- Bilabial (b,p) consonants with central vowels (a)= neutral tongue
- Frontal (t, d) consonants with front vowels (i)= tongue stays front
Chain shift
Substitute a different sound (guck for duck)
Stored articulatory routines
Can’t generalize sounds to other contexts
Syllable Simplification
Cluster Reduction- stop to top
Final Consonant Deletion- bat to ba
Reduplication- doggy to dada
Weak Syllable Deletion- banana to nana
Fast Mapping
“Give me the chromium crayon” then repeat a week later
-mapping between sound and representation
Original word learning theory
people show kids objects then tell them the name
Word learning constraints
Biases that limit the hypotheses that the child entertains
Taxonomic Constraint (word learning constraint)
- Assume words refer to category of things that are the same
- Assume words don’t refer to category of things that have related associates
Cake—> other desserts, other candles, NOT kinds of parties or kinds of lights
Extension of novel words- categorization, “dax”
Whole object constraint (word learning constraint)
- Assume that word refers to entire object
- Don’t assume that word refers to part of object, its substance, its color
Whole object bias helps us learn nouns but slows verbs, adjectives, etc
Mutual Exclusivity (word learning constraint)
- Assume every object has one name
- Assume there are no synonyms
- Can override whole object constraint
Social cues in word learning
Word learning occurs socially, child is attempting to discover another’s intent to refer
Joint Reference
When speaker and listener both interpret a phrase as the same thing
Follow-in labeling
Mothers label the object a child is looking at
Disjoint labeling
A parent isn’t talking about the object the child is looking at
Why do children learn nouns before verbs
3 hypotheses:
- Input determines word learning- children learn words that parents directly teach
- Limits of cognitive development- infants don’t have abstract concepts/words
- Limits of linguistic development- relational words relate concrete words
International Adoptees
Lose native language and switch to new language, by adulthood no evidence of native language.
- verbs increase as vocab does
- no difference in vocab composition in adopted vs infant learners
- shifts from nouns to verbs are not cognitively driven
- adopted children actually learn words faster
- vocab and syntax work together
Children’s morphology
- smallest unit that has meaning
- start to appear in two-word stage
- first acquisition is present (ing) and last is auxiliary contractable (she’s going to school)
- tend to over regulate past tense and plural
How to children learn morphology? (theories)
- They memorize
- They analogize
- They go by rules
Linear syntax
pronoun cannot precede main referent
Co-referential syntax
pronoun cannot c-command main referent
-children know principle C
C-command
Node B commands node __ only if:
- B does not dominate __
- ___ does not dominate B
- The first branching node that dominates B also dominates ___
