AI summary by TwinMind
Lecture 25: Language Acquisition
MIT OpenCourseWare
58m 15s
Review of Binding Theory and Pronoun Reference
- Reviewed semester opening examples: "Who did Biden ask to see him?" vs "Why did Biden ask to see him?"
- First example: "him" can refer to Biden (outside bracketed domain)
- Second example: "him" cannot refer to Biden due to PRO binding
- Student Faith noted wh-movement from islands concern - professor confirmed students now "warped" by syntax training
- Explained complex pronoun reference patterns through:
- Wh-movement of "who" from object position of "ask"
- Null PRO subjects in infinitival clauses
- Object control vs subject control with verb "ask"
- Principle B: pronouns must be free in local domain
Plato's Problem in Language Acquisition
- Central question: How do children acquire complex grammatical knowledge without explicit instruction?
- Evidence against parental teaching:
- Actual parent-child dialogue examples showing parents focus on truth, not grammar
- "Mama isn't a boy he a girl" / "That's right" - mother ignores pronoun error
- Failed correction attempts: "nobody don't like me" dialogue repeated 6 times unsuccessfully
- "Give me other one spoon" correction attempts fail despite word-by-word instruction
- Universal Grammar hypothesis:
- Humans preset to build languages in certain ways
- Need data to determine which specific language being learned
- Core grammatical principles (Control, PRO, Principle B) are universal defaults
Critical Period Evidence
- Anecdotal evidence: Adults struggle with second language acquisition compared to childhood ease
- Case studies of language deprivation:
Genie Case (discovered LA 1970)
- Isolated first 13 years: strapped to chair, punished for sounds, no linguistic input
- Father paranoid schizophrenic who hated noise
- One sibling died from exposure in garage
- After rescue: learned individual words but never grammar/sentence structure
- Scored normal intelligence on tests but never developed syntactic abilities
Chelsea Case (Northern California)
- Parents believed she was developmentally disabled
- Actually profoundly deaf - discovered at age 12-13
- No exposure to spoken language or sign language
- After cochlear implants: learned words but not sentence construction
- Happier case than Genie (loving parents, no abuse)
- Now better screening prevents such cases
Counter-evidence: Hyperpolyglots
- Ken Hale example: MIT professor who learned languages extremely rapidly
- Spent 1960s in Australia learning Aboriginal languages
- Could become fluent in one week
- Some communities complained later linguists were "slow" compared to Hale
Experimental Methods in Language Acquisition Research
Sucking Rate Tasks
- Infants suck rapidly initially, then slow down
- Speed increases when presented with novel/interesting stimuli
- Japanese infants distinguish R vs L sounds (adults cannot)
- English infants distinguish aspirated vs unaspirated stops
- 50 millisecond cutoff for aspiration distinction
- Theory: babies born ready to distinguish all possible sound contrasts, then lose unused distinctions
Speech Segmentation Research
- Children make segmentation errors: "I am being have" (from "behave"), "your Amy" (from "Miami")
- Use statistical learning for word boundaries
- Saffran et al. (1996) experiment with artificial CV.CV.CV words
- Children learned transitional probabilities quickly (e.g., if hear "to" expect "la" next)
- Performance varies with syllable number and prosodic cues
- Result: 13,000 words by age 6 (6-7 words/day learning rate)
Syntactic Development Patterns
Japanese Child Phonology (3;2 years old)
- Changed K→T in all contexts (mikang→mitang, poketto→potetto)
- BUT maintained adult Japanese T-allomorphy rules:
- T→tsu before /u/ (matsu not *matu)
- T→chi before /e/ (machi not *mati)
- K-derived T did NOT undergo allomorphy (aki→ati not *achi)
- Shows underlying K/T distinction despite surface neutralization
Root Infinitive Stage
- Common cross-linguistically (Danish, Dutch, French, German)
- Children use infinitive verbs in main clauses where adults require finite verbs
- German example: V2 vs verb-final positioning
- Finite verbs correctly placed in second position
- Infinitive verbs correctly placed clause-finally
- Andreas (German 2-year-old): 281 utterances analyzed
- Finite verbs mostly V2 position
- Infinitives mostly clause-final position
- Children confused about infinitive licensing but not verb placement rules
Experimental Methods: Truth Value Judgment Tasks
- Puppet methodology: easier for children to correct puppet than adult
- Quantifier interpretation differences between adults and children
- "Every rabbit is riding an elephant" experiment:
- Picture: 3 elephants, 2 rabbits (rabbits on 2 elephants)
- Adults: "Yes" (every rabbit is indeed riding)
- Children: "No" (point to elephant without rabbit)
- Children appear to have bidirectional quantifier interpretation
- Active research area investigating quantifier acquisition patterns
Research Methods Summary
- Naturalistic recordings: CHILDES database of transcribed child speech
- Consultant work with children (like field work but with cookies and crying)
- MIT lab conducts this research, seeks volunteers
- Undergraduate language acquisition course available
