word learning Flashcards
learning
morphology can give a clue to meaning - can use existing knowledge to work out
-vocab knowledge strong predictor of later outcome
vocab + reading comprehension are best predictors of school outcomes.
early phases
- Pre-linguistic - 9 months - crying, going
- holophrastic - 1 year - one word = many meanings
- telegraphic - 18-24 - 2/3 word
- stage 2 grammar - 2-3yr - plurals, overgeneralisation
- adult like speech - 5-6 yr
when learning new words need to discover:
- relevant linguistic units
- relevant conceptual units
- establish mapping between them
skinner - conditioning
children learn solely from inout
- when corrected - what do parents focus on
- parents reinforce semantic correction
- other times correct semantic but use child language
- limits of input
- have to wait till child is read
- pushing child can be good but no point if they arent reads
Chomsky - LAD
grammar + syntax too complex - born with universal grammar/syntax
sorts words into syntactic categories
-need to consider extent cognitive processes are language specific
6 months old:
- mommy - looks at mum video
- daddy - looks at dad vid
- understanding always proceeds production
- not when presented unfamiliar person
comprehension before production
golinkoff - 17m old looks at correct picture - 1 word stage can understand wod order
vocabulary spurt
occurs once reach 50 words
increasing memory + cog capacities
vocab size + when children start 2 word utterances is strongly related
66% of vocab is object words
guessing the meaning - gagagai
give the unknown name to something that doesnt already have a name
whole object constraint
nouns > verbs
verbs occur in lots of concepts + happen under time
before can understand a verb, have to know about the nouns involved
over/under extensions
over - bird = birds
under - duck = only toy duck in bath
whole object constraint
word refers to whole objects rather than parts of objcts
adults clear when not referring to whole
-may reflect more general bias to divide the world into discrete objects
processing constraint - innate bias to group objects and perceive them as one thing
mutual exclusivity bias
each object can only have more than one label
shape - objects of same shape have same name
taxonomic constraint - assume similar category objects have same label
-as form new words - new categories - new category = new words
syntactic bootstrapping - good at perceiving if a word is verb/noun based on other cues
problems with constraints on learning
not necessity
could make llearning harder
where do they come from
are they only at certain points of development?
explanations of constraints
associative/attentional learning
- perceptual cues co-occur in predictable way
- attention selective for those cues
- shape bias
lanuage constraints emerge from general learning biases
probability theory - statistical learning - domain general
fast mapping - Carey + Bartlett 1978
new word connected with concept after only brief encounter
- use context to make guess about maning
- novel object quickle ass by 15-2 year olds - remember new term when tested later
-retention 1 week later = fast mapping of novel name to novel object
Horst
24m olds - good intiial learning
- but couldnt do it 5 mins later when presented with other novel objects
- not exactly learning the new object/word mapping
- good way to test if theyve learnt what the word represents or just he word in the context
fast mapping
sleep crucial for integrating new words - role of memory
Gathercole - speed at which 4 year olds can recall non words correlated with current vocab + growth
pragmatic and social factors
motherese (CDS) - helps stand out in speech
social factors - taking turns
parents establish shared attention
relies on memory capacity - infantile amnesia
related to development of concepts
private speech - later inner speech
how do we add new words to the lexicon?
the cohort model - Marslen -Wilson
-as the speech unfolds the words accessed become more limited till found uniqueness point
novel word learning - need to hear more word to recognise - UP has shifted
Dunmay 2007 - test
look for lexical competition effects straight agyer learning and after delay
-whether processing of existing point is lower - slows consolidation and integrated - engaging lexical competition
- new effects absent immediately but present after sleep
- findings extended to children
how do we integrate these new words?
2 seperate compoents:
-hippocampus - plasticity - fast learning of new word
-neocortex - stability - lt storage of lexical info
why not integrate new info straight away?
computational property of the brain
- advantages to representing in distributed + overlapping form
- disadvantage - catastrophic interference
would we expect the same mechanism early in development?
maybe - delayed improvements in recognition for new words
maybe not
- different sleep
- less existing knowledge/competition
- heightened plasticity
SIMILAR IN CHILDREN
- AM + PM group
- word learning greater in PM group when sleep
- AM group - when had slept the next day, similar level to PM group
