the acquisition of our first language part 2 Flashcards
what is the main problem when discriminating words as an infant
the segmentation problem
what is the segmentation problem
-fluent speech does not provide unambiguous indications of where words start or end
-speech doesnt have the equivalent of spaces between words
-learning lang involves figuring out which sounds clump tog to form words
-parents only to speak to infants in one word utterances 10% of the time
-babies are often confronted with many words sewn seamlessly tog and must work out where the edges of words are
what are the techniques to solving the segmentation problem
1.isolated word hypothesis
2.phonotactic constraints
3.the prosodic bootstrapping hypothesis
4.infant directed speech
5.statistical learning- transitional probabilities
what is the isolated word hypothesis
-1st technique to resolve segmentation issue
-bortfeld et al 2005: babies use this at 6 months ish
-baby captures strings of sounds within a sentence
-the sounds regularly heard in isolation must be a word
-to fill in the rest the baby must make inferences
what is phonotactic constraints and what is a weakness of it
-2nd technique to resolve segmentation issue
-sensitivity at age 7 1/2 - 9 months
-uses intuition of which sound sequences are allowed at the beginnings and ends of words e.g there are no english words starting with pk/pt etc
-baby might reject the word ptangb bc of ingrained templates which this word does not match
-weakness: if infant does not know where word begins or ends, how will they learn some sequences are more frequential at the beginning or end or a word
what is the prosodic bootstrapping hypothesis
-3rd technique to resolve segmentation issue
-use of stress patterns to segment speech
-90% of bi-syllabic words have stress syllable at start of word (trochaic)
-trochaic e.g BOttle, iambic e.g boTTLE
-weber 2004
–stress patterns not seen in infants at 4 months
–at 7 1/2 months infants recognise words with trochaic stress patterns
-errors are common here e.g guiTAR is might be put as one word TARis until 10 months when errors stop
-older infants do not rely on stress patterns as much
what goes on when babies solve the segmentation problem
they begin to understand words have meaning
what is statistical learning- transitional probabilities
Saffran et al 1996
-5th technique to overcome segmentation issues
-babies exposed to stream of sounds from invented language for 2 mins
-sound stream could be segmented in many diff ways
-infants quite capable of segmenting words even w artificial language
transitional properties
-brain knows /tee/ follows /pri/ 80% of time so brain assumes these as a package
-high probability pairs become words and low prob pairs treated as separate speech entities
what is infant directed speech (IDS)/ motherese
-4th technique to resolve the segmentation issue
-exaggeration of the prosody of speech and is short
-not all cultures have IDS
what is the main challenge of understanding words have meaning
poverty of stimulus
-environment does not provide enough info for child to identify 1 meaning of a word so often errors occur
what comes first understanding or production
understanding (perception before production)
what is it called when an infant learns a word they have only heard once
fast mapping
when does understanding language occur
-end of 1st year: initial associations of sound sequences and meaning
-understanding is dependent on context and is a gradual process
-10months: 40 words understood
-18 months: 250 words understood
process of speech production
-12 months: produce 1st word
-12-16 months: 2/3 words learnt per week
-18-22 months: vocab of 50 words
-22 months +: word spurt = 8/9 words per week
does this fast learning of language support the LAD
yes, indicates a special mechanism for word learning (predisposition for lang)
markson and bloom 1997 study
-found children and adults have similar abilities to learn linguistically
-pp presented with facts about objects given only a few exposures
-visually presented info not retained as well
how is it a child knows that the word rabbit refers to the whole animal and not the fur
-whole object bias in word learning
-when babies hear a new word in the context of a salient object, they likely assume the word refers to the whole object not the colour or surface
stager and werker study 1997
-infants learn nonsense words that are minimal pairs (differ by one phoneme) e.g bih vs dih compared against words that are not e.g lif vs neem
-habituation phase: dih, dih, dih etc associated to a toy
-test phase: toy initially dih changes to bih
-toddlers age 14 months or under cannot distinguish between the 2 words
what is the mutual exclusivity bias
16 month olds can learn a new word for an object but not if the object already has a name
why can toddlers not distinguish between the words bih and dih in stager and werkers study
-we are tight are what we commit to memory
-if words are very similar we might assume they mean the same and only remember one
-inefficient retrieval
what are the 3 processes in the acquisition of word meaning
-word labelling
-categorisation process
-network building process
what are errors evidence for
how children learn meanings and categories
what is under extension
-word in very specific way e.g cold milk, cold only used to describe milkw
what is over extension
-gap filling, mental fog, wrong analysis
-e.g all animals with 4 legs are dogs
how do children create their own categories
based on prototypes
what indiv diff is the acquisition of vocab subjected to
gender, birth order, quality and quantity of surrounding lang, processing abilities etc
