ACQUISITION I: Learning word meanings
Tension
Whether words are learned gradually i.e. incrementally by accumulating evidence across several situations in which the same phonological segments occur
or if you make associations on the first encounter
Behaviourist theories of learning
Learning was continuous and studies showed a similar learning curve
ROCK 1957
control condition > participants were given a set of 12 letter-number pairs which were learned to criterion
experimental condition > a pair was removed and replaced if the participant failed to recall it
Results Rock found that there was no difference between the experimental and control groups, indicating that ‘learning occurred not with a gradual, incremental increase in strength of memory traces but rather in an all-or-none fashion’
CRITICISM
Item selection effect: by experimental design, the more difficult pairs were removed
Challenged traditional behaviourist models of learning
REPLICATIONS
Clark et al (1960); Estes (1960) obtained similar results
click to edit
Useful info
Quine (1960)
The PoS argument, rarely a direct association between language and the referent
Williams (1961)
Extra control condition where PP's studied unrecalled pairs in the EC
found that the control group that learned the experimental group’s eventual list took significantly fewer trials to reach criterion
suggest that was an IS effect so Rock (1957) cannot be used as a evidence for all or none learning
Fast Mapping
Medina et al (2011)
Coined by Carey and Bartlett (1978)
account for the limited information which is gleaned from a first encounter with a word, but provides the framework to allow later information to be added. mapping is extended after several encounters
tested acquisition of a new colour word chromium associated with olive green in pre-school age children and their knowledge that chromium was a hyponym of colour
Results
Many children struggled with the hyponym class, suggesting unstable lexical relations at this age, although half of the participants picked up something about the word chromium or the naming of olive from a single exposure to the word = fast mapping had taken place
later tests of the children’s performance with chromium, once they had had more exposure to it, show a marked increase which indicates that increase in exposure aids in word acquisition.
Yu and Smith
2007
2008
They asked adult learners to simultaneously learn relatively many word-referent pairs (18 at a time) from individual learning trials that were highly ambiguous.
In all conditions, the participants learned more pairs than would be expected by chance.
suggests that participants ‘calculate cross trial statistics with sufficient fidelity and by doing so rapidly learn word-referent pairs even in highly ambiguous learning contexts’ (Yu & Smith, 2007:414 )
In the learning phase12- and 14-month-old infants 2 objects were simultaneously presented followed by 2 words
non-verbal, infants are capable of distinguishing word meanings probabilistically after having been exposed to them, in an ambiguous environment, multiple times
could suggest non-innate
click to edit
Criticism
this experimental method not representative of real world acquisition e,g, rarely are children only presented with a handful of objects in view
using several identical images also a distortion of reality since learner almost never encounters referent in exact same context
Statistical model assules learners hold several possible meanings in their head at once whihw would place significant burden onthe working memory as it would have to do this for every word
Gavagai
Mattys & Jusczyk, 2001
infants are able to segment units of running speech by relying on a variety of clues including prosodic features and statistical regularities