all fla

Stages of Acquisition PBPLGD

1. Perceptual dev - rhythmic awareness

2. Babbling - motor exploration

3. Phonological dev - building sound inventory

4. Lexical dev - word boundaries/meanings

5. Grammatical dev - combining units

6. Discourse - managing convos, intention and pragmatics

Nature Theory

• innate dispositions

• newborns attuned to cues for learning

• lang is species-universal and specific

- CHOMSKY

Nurture Theory

• learning requires exposure to lang

• environment, not genetics

• no exposure = no ability (Morford 2002)

The Brain

• two sides (Left hem/Right Hem)

• sides are contralateral (LH controls right side etc)

The Brain + hemispheric lateralisation = RH

• sensorimotor - left half of body

• visual-spatial processing

• global processing

• responsible for - pragmatics, prosody

The brain + hemispheric lateralisation = LH

• sensorimotor - right half of body

• analytic, detail-oriented

Home of lang centres:

• Broca’s area - lang production

• Wernicke’s area - lang comprehension

Early Development

• babies begin to develop speech perception

• prosody, sounds, visual info

First Perception of sound - Mehler et al 1988

• 4 day old infants attention to their L1

• babies can discriminate between sounds of native vs everything else

• via sound waves and rhythm

Visual cues in speech processing

McGurk effect - audio/visual doesn’t match, so we assume visual

• children watch speaker’s articulation and collect info

Types of babbling (4)

1. cooing - vowel sounds - 2/4mo

2. exploratory - shrieks - 4/6mo

3. reduplicated - repeat consonant/vowel structure (bababa) - 6/9mo

4. variegated - diff vowels and consonants - 10/12mo

Speech Segmentation

• finding boundaries between words

ambiguous segmentation

• oronyms - homophonous phrase = sequence of sounds consistent with more than one sequence of words

— i scream and ice cream

Segmentation cues - stress

Culter and Butterfield 1992

• lexical words study

• only 8% words have weak initial syllable

Segmentation cues - statistical learning

Saffran et al 1996

• infants quickly learn about syllables occurring together in speech, listening longer to ‘nonwords’ as unfamiliar

• 2min exposure to made up lang= 8month olds learn where word boundaries are

lexical challenges for children

• learning to produce sounds systematically

• learning the meaning of words

Jakobson 1941 - Maturation model CECPMP

Jakobson 1941 Maturation model

• 6 elements of verbal comms

1. context - message’s context and reference to world

2. emotive function - speaker attitude, emotion and expression

3. conative function - receiver and aims to be influenced by speaker

4. phatic function - prolonging/interrupting communication channel

5. metalingual function - code, ensuring sender and receiver understand

6. poetic function - message and its feel, aesthetic forms

Explaining phonological errors

Berko and Brown - Fis phenomenon

• production/comprehension lag

Leopold 1970

• children sometimes develop backwards correct pronunciation replaced by mispronunciation

—

• children often delete unstressed syllables, giraffe > raffe

Vihman et al 2014

• errors occur when infants apply templates for existing words to new words

— pretty > biddy

— piano > pinano

Nouns-first bias

Getner 1982

• 1,982 nouns salient to children

• need to acquire first crosslinguistically

Word learning and meanings - theorists

Tomasello 2000

• children good at understanding context and speaker intentions

Baldwin 1991

• joint attention by 9mo, children and caregivers coordinate visual attention

• Harris et al 1984

• caregivers attend to child’s gaze, narrate what child is attending to first

Two-word stage - transition

Goldin-Meadow and Butcher 2003

• word + gesture instead of word + pointing

• more holophrases = night-night

• vertical constructions = more milk, car outside

Learning grammatical morphemes SPFC

Brown 1973

• acquired in similar order by children

based on said factors:

• salience - phonological stress/perceptibilty

• predictability - consistency in language

• Frequency

• Conceptual complexity

Imitation and Novelty

• Skinner 1957 (Behaviourism) - lang acq is from echoics + reinforcement

• Chomsky 1959 (Nativism) - part of lang is imitative, but goes beyond imitating specific utterances (innateness+)

development oof children learning past tense

Cazden 1968

1. child produced bare forms as past tense = i make

2. learns irregular forms correctly = made, went, did

3. learns regular -ed verbs = walked, played

4. forms generalisations about past tense > overregularisation = goed, eated

5. child figures out exceptions

Acquiring past tense English

Two models:

• Dual route = words and rules model

• Single route = analogical/connectionist

acquiring past tense English - Dual route

• learning specific to items (irregular then regular)

• Pinker 1999 = overgeneralisation makes learner deduce regular -ed rule

• irregularity can spread

acquiring past tense english - Single route

• words learned by route

• patterns detected between similar verbs - sleep/slept

• regular verbs strongest pattern

Rumelhart and McClelland 1986

• connectionist networks

• single mechanism drives all learning based on individual items in memory

What approaches explain how children learn grammar?

Nativist approach

• born with lang-specific mechanisms

• mental ‘switches’

Usage-based approach

• born with general learning mechanisms to build up grammar based on heard examples

• ‘islands’ of development

Broad vs Specific views on innateness

• species-specific

• species-universal

— children learn complex grammar despite varied upbringings

Nativist approach to grammar acquisition

• argument for poverty of stimulus = linguistic input encountered cannot account for linguistic knowledge developed (Chomsky 1980)

Children acquire lang even though input:

• contains disfluencies and incomplete utterances

• contains no negative evidence

• finite

Poverty of Stimulus

• claim 1 - children at age x know y about language

• claim 2 - these children do not have sufficient language experiences for them to know x

Argument against poverty of stimulus (nurture)

• children acquire lang through general pattern recognition rather than reliance on innate dispositions

• not direct mirroring

Proposals for innate principles

Children born expecting

• sentences have constituent structure

• particular constituents (NPs/VPs)

• syntactic units have head/complement structure

• heads before complements or complements before heads

Basic Constituent Order

• English = SVO (subject verb object)

— The dog ate the ball

• verb precedes object

• preposition precedes nouns

Usage-based model of grammatical acquisition

• Nativist view = impoverished data, parameter setting, universal grammar

• Usage-based view = grammatical development proceeds slowly based on examples learned case-by-case

Akhtar 1999 - Children willing to consider different constituent orders?

• studies english speaking children in 3 groups (2;8, 3;6, 4;4)

• SVO - “look, Elmo dacking the car” = new word

• SOV - “look Elmo the car gopping”

• VSO - “look, tamming elmo the car”

— broad generalisations entrenched by 4 (replicated with younger children by Abbot-Smith et al 2001)

• 80% match

— challenged parameter-setting model

Akhtar 1999 - Item-based learning

• many children learn frames for specific verbs

• eg suffix -ed for past tense

— kick to kicked

• No immediate SVO generalisation

• Tomasello 1992 = Verb Island Hypothesis

• any model of grammar requires generalisation at some point

Item-based Learning 2

Lieven et al 1997

• studies of parents and children’s speech show most utterances (2-4yrs) follow set patterns in adult speech

Lieven et al 2003 - case study

• Annie age 2

• 63% utterances match previous utterances in corpus

• 74% novel utterances = almost exact match previously via adding or removing word

MacWhinney 2004

• Innateness

• Children born with useful capacities for lang

• Conservatism - generalise cautiously from input

• Correct errors with various skills

— diff mechanisms can compete (memorisation vs generalisation)

— correct output when similar

Domain-general cognitive skills

• Lang may be innate yet involve no mechanisms that evolved specifically for lang itself = Brown and MacWhinney 1989

• Various cognitive skills - like statistical learning and pattern recognition, have been recruited by lang but have kept their ‘day jobs’ in other areas of cognition = Elman and Bates 1997

Hartshorne et al 2018 - critical periods

• international study (670,000 pp)

• checks ccuracy on 95 items in grammar

—

• learning rate drop at 17

• lang exposure by 12 is best, cp closes 17

Age and acquisition - nativist approach

Nativist

• strict critical period

• lang exposure neded but capacity biological

• mechanisms innate (LAD)

Age and acquisition - nurture/usage approach

• age effects gradual = sensitive perod

• neuroplasticity in children

Hockett 1960 language design features CPAD

1. communicativit

2. productivity

3. arbitrariness

4. duality of patterning

Arbitrarness in sign

Kilma and Bellugi 1979

three levels:

• transparent

• translucent

• opaque

Duality of patterning HPMO

Battison 1977

phonology in sign includes 4 parameters

• handshape

• place of articulation/location

• movement

• orientation

L2/late vs L2 English

Mayberry and Lock 2003

1. native

2. L2 English

3. L2 English with bilingual L1

4. Late learners

ASL late L1 vs L2

Mayberry 1993

• comprehension problems for late L1

• Late L2 better

Imperfect Learning Model

Pinker 1999 - lang not acquired perfectly

• results in inaccurate reproduction and imitation

Problems with imperfect learning

• frequent phrases undergo phonetic reduction (i don’t know to i dunno) = Bybee 2007

• Correlation between frequency and innovation with L1 exposure = Bybee/Beckner 2010

Holistic acquisition

Nelson 1973

• picking up on langauge through phrases and ‘chunks’ as single units

• ‘whatisit’ / ‘idunno’ as single units

Analytic acquisition

• building vocab word-by-word

• early lexicon dominated by nouns

Syntactic bootstrapping

Gleitman 1990

• use structure of unfamiliar words to deduce meanings

• usually verb learning

• example of wug test

Semantic bootstrapping

Pinker 1984

• meaning first, then grammar

• children acquire grammar by learning meanings and then inferring grammatical structure

• linking rules

Prosody - in earliest receptive learning and speech production

— Melodic line of speech

• first windows babies have into structure of language

• in womb via rhythmic sound waves

Categorical Perception (children)

— Brain groups continuous sensory inputs (like sounds/colours) into categories

Children

• 6-8m discriminate subtle phonetics

• 3-6y establish categories for speech sounds and facial expressions

Categorical Perception (adults)

• categories become sharper - refined and consistent

• better at ignoring within-category variations (diff versions of letter d)

• can process words automatically

Habituation/Dishabituation

• H - decrease in response to a repeated stimulus (tune out)

• D - rapid recovery of original response, triggered by introduction of sudden or new stimulus

— live by airport and don’t hear them anymore BUT storm wakes you up and hyper-aware of planes again

Lack of invariance problem

• no 1-to-1 unchangeable RS between specific speech sounds (phonemes) and its representation

• due to coarticulation or speaker difference (accent)

Visual attention in development (4)

• how infants and children learn to select, process and focus on visual info in their environment

1. Newborn phrase (involuntary) - babies drawn to high contrast movement and bridghtness

2. 2-6m (emerging cortical control) - process finer detail, attention from external features to internal facial features (eyes/mouth)

3. 6-12m (joint attention) - endogenous control and focus on toys/tasks longer

4. Early childhood to adolescence (exec control) - learn to ignore distractions and focus on tasks for longer 9selective vs sustained attention)

Speech segmentation

— process of identifying boundaries between words, syllables or phonemes

• listeners utilise stress patterns and pitch changes

• tracks probability of syllables occurring together and sound combinations

— ie in another language

Statistical Learning

— detecting implicit probabilities in sound stream

• infants show they can isolate words based on frequency and probability

• challenges innate ideas

Comprehension vs Production

C - what is heard or read

• starts in womb

• 6-9m basic word recognition

• Age 1 = dozens of commands but produce ½

P - Speaking of Writing

• lags behind comp

• first words around 12m

• combining words 18-24m

Template model

• theoretical frameworks where children implicitly use structural templates to map out and produce complex language

• nature approach - social interactions

Noun-first bias and controversy

Macken 1995

• learning concrete objects and people quicker than verbs/adjectives in early vocabs

• controversy = is bias innate or byproduct of languages spoken to them?

Fast-Mapping

— children rapidly acquire/retain meanings of new words after minimal exposure

• builds ‘skeletal’ placeholder for its meanings and refines over time and exposure

Referential Indeterminacy (Quine 1960 Gavagai)

• No objective fact of the matter regarding exactly what object or concept a word’s meaning is from

• Rabbit could mean verb, part of the animal etc

Whole object assumption/bias

• children naturally assume new word refers to entire object than its individual parts, properties or actions

— ball could mean to kick, the colour, the shape, etc

Mutual exclusivity assumption

• where children assume that object can only have one single name

— if child knows that dog is and hears adult point at unfamiliar dog and say wug, child assign’s wug to the unfamiliar animal.

Joint attention & social-pragmatic accounts

JA

• coordinate attention between person, object and other person - 9-12m

SPA

• Bruner/Tomasello

• children learn how to use language by understanding communicative intentions of others

Telegraphic stage

• infants 18-30m speak in short, simple, 2-3 word phrases

• primarily using nouns, verbs and adjectives

— Daddy go

— Where ball

Holophrastic Speech

• single word to express entire idea

• 9-18m

• heavily relies on context and tone.

— food

— ball

— up

‘Wug Test’ Berko 1958

• children generalised grammatical rules rather than memorising words

• experimenter asks child to pluralise a ‘wug’ creature

• supports nativism - brain wired to apply rules when needed

Factors in learning grammatical morphemes

1. semantic complexity - how easy concept is to understand

2. Input frequency - often its heard

3. perceptual salience - noticeable in speech

4. native language transfer - for L2

Over-regularisation

• children apply regular grammar rules t irregular words

• goed/went

• tooth/teeth

• sign of cognitive development

U-shaped curve

• non-linear learning where performance starts high, temporarily declines, then rises again

1. High accuracy - children memorise through imitation

2. Dip - overgeneralisation (goed/went)

3. Target performance - figure expectations to rules and synthesise learning

Poverty of stimulus argument

• nativist - Chomsky 1980

• children acquire lang to quickly for it to be explained solely by environment

• environment too limited

Parameter-setting model

• Humans born with universal grammar principles and binary switches (para).

• Child acquires native lang by setting mental switches baed on what they hear

Verb island hypothesis

Tomasello 1992

• young children don’t initially possess abstract grammatical rules

• learn verbs as isolated ‘islands’

• cannot grasp syntactic categories

Simultaneous bilingualism

• when child acquires two native langs concurrently from birth/very young

• relies on neuroplasticity

Sequential bilingualism

• child acquires second language after establishing their first lang

bilingualism and lang discrimination

• glottophobia = unfair treatment or prejudice of individuals based on lang/accent/vocab.

• bilinguals = biases against non-native accents or code-switching

bilingualism and vocab size

• while bilinguals have smaller vocabs in single lang, their total vocab across both languages is equal or larger to monolinguals

• grasping underlying concept of words

Critical vs Sensitive period

• C - strict, bio determined window of time when environmental input is required for normal acquisition, consequences suggest irreversible defects

• S - Flexible, gradual window whereby language development requires inputs, consequences suggest harder learning

Imperfect learning model

• L2 acquisition

• difficulty fully internalising certain features of a non-native language

• missing native-like mastery of specific grammar features

Spatial Modulation (signed langs)

• 3D signing space to represent grammar, establish characters and convey RS between nouns