linguistics- language learning

what do infants need to learn and why is it challenging

• they need to learn to distinguish distinct phones in word from allophones, therefore distinguishing meaning from variation

• they need to perform word segmentation which is challenging to pick out word boundaries from continuous streams of speech

how do babies perform word segmentation and how is this investigated

• it is investigated in the head-turn paradigm where babies heads turn in response to familiar sound sequences

• they use familiar words as anchors and are used to detect where neighbouring words and boundaries are 

what else are babies good at that helps them in learning language 

• they pick up on ‘phonotactic constraints’ which are rules for a certain language on which sounds can occur together in combination

• infants are good ‘statisticians’ meaning they are good at tracking transitional probabilities between sounds which can be used to infer word boundaries

what is the old vs new lego analogy

• old legos represent empiricist ideas of domain generality, as there are lots of small identical parts that under lots of effort can be combined to construct something new, though this requires major environmental input

• new legos represent nativist ideas of domain specificity, as there are a smaller number of more complex pieces that have been designed for a specific purpose, minimal assembly is required but the pieces (ideas) are harder to reuse 

• these represent radical and extreme points of each view

what did Chomsky argue about empiricism

• a strong opponent of empiricism

• argued that language and knowledge too complex to be explained by experience alone so must be supported in learning by innate mental structures 

what did Skinner argue on empiricism

• represented radical behaviourist mindset

• argued that internal processes should not be treated as mental phenomena, but should be understood the same way as external behaviours and through objective, observable data 

what did Elman argue on empiricism

• Elman was not a radical empiricist and bridged the gap between the two traditional sides of the debate 

• he rejected ideas of innate mental structures but allowed for innate biological constraints like consistency in developmental timing 

• he proposed that complex cognitive abilities emerge from the interaction between innate neural structures with experience and learning 

are empiricism and nativism enough to account for language learning ?

no because development of language is a complex interaction of children’s inner capacities and their environment 

do adults find word segmentation difficult

yes, shown through the difficulty in segmenting words when hearing a foreign language and also in the mishearing of song lyrics 

who researched the segmentation of words by children and adults

Jenny Saffran

how did she propose babies learned to segment words

• by working out and tracking the probabilities of sounds occurring together

• if probability is high then sound is likely word internal, if probability is low then more likely to mark a boundary between words 

• this can be modelled on training data

how does Saffran’s data support empiricist accounts

it demonstrates a reliance on domain generality as tracking probabilities arises from environmental experience and applying a general analytical not linguistic skill

how does Saffran’s data support nativist accounts

a nativist might argue that the ability to track probabilities refutes the blank slate idea as it is still an innate ability, however strict nativism argues that these abilities are strictly linguistic to help learn language