statistical learning

Question 1

can infants learn about language in the womb?

Answer 1

yes - they can hear language, patterns of speech - not innate

Question 2

do infants make contrasts with other languages while in the womb?

Answer 2

• they only make contrasts present in their own language while in the womb - new borns are better than 1 year olds - they differentiate between lanuages

Question 3

example of infants only making contrasts in their own language -

Answer 3

in Hindi 2 ‘d’ sounds. this is discriminated in all newborn infants and Hindi 8 month yr old but not English 8 month old.

Question 4

infants are born with the capacity to learn languages but…

Answer 4

if they aren’t being taught it they lose that ability

Question 5

Chomsky - learning theory

Answer 5

argued that parents reward children for
grammatical utterances, punish them for ungrammatical utterances - BUT this does not happen

Question 6

2 more things that Chomksy believed?

Answer 6

• believed certain aspects of language
  (e.g., grammar) were innate because of the poverty of the stimulus (the
  input available to the child was insufficient to be learnable)
• believed in a universal grammar - all languages share certain principles - innate

Question 7

research disproving some of chomsky’s beliefs…

Answer 7

research can’t identify grammatical structures common to different languages

Question 8

what do learning theorists argue today?

Answer 8

-that language can be learned using general learning mechanisms - like statistical learning - learning of patterns

Question 9

language is -

Answer 9

predictable

Question 10

how is language predictable?

Answer 10

gramatical categories occur in predictable positions in sentences = frequent frames
these frames occur reliably in child input… help children learn about grammar (Mintz)

Brain is specialized for statistical learning - which assists language learning

Question 11

brain is specialised for…

Answer 11

statistical learning which assists language learning

Question 12

what did that recording show us
(four “words”: tupiro, golabu, padoti, and bidaku)

Answer 12

stimulus is not impoverished at all- the info is there in the environment already

also showed that babies understand the word even if they don’t comprehend it

Question 13

overview of the method of Teinonen’s study

Answer 13

1. sleeping newborns - listened to nonsense sound.
2. wakes up and listened to same sound with new words added.
3. measured brain activity

Question 14

results of Teinonen’s study:

Answer 14

brain responds differently to old and new words

pattern recognition is excellent - even while asleep they use familiarity to group together sounds/ words.

Question 15

behaviour is predictable - 2 different theories

Answer 15

• infants have an innate theory of mind (understanding that people have goals etc)
• infant’s statistical learning allows them to learn about pattern behaviour

Question 16

Baldwin’s experiment

Answer 16

infants shown videos of complete and non-complete actions. they are more interested in the mid-action sequence.

Question 17

interpretation of Baldwin’s experiment

(hint - infants )

Answer 17

• infants do not need an innate social understanding
• infants learn to predict patterns in behaviour
• only years later do they work out there are mental states behind behaviour

Question 18

so what is innate?

Answer 18

statistical learning - the ability to learn about patterns in language and behaviour - etc

Question 19

SLI:

Answer 19

special language impairments
- short sentences - small vocab etc

Question 20

3 conclusions of Evans et al:

Answer 20

Typically developing children are good at learning about patterns in syllables and tones

Children with SLI have general difficulties learning about patterns in syllables and tones

Better statistical learning = better language ability

Question 21

question of Evan

Answer 21

do SLI children have a specific difficulty with language stimuli or general difficulties w statistical learning

Question 22

statistical learning with autism - hypothesis

Answer 22

would children w autism have difficulty detecting co-occurences amongst syllables ?- as they have poor language and understanding others ability

Question 23

studies that have examined activity in brain areas in certain brain regions when listening to random / nonrandom syllables show…

(autistic children vs typically developing children)

Answer 23

typically developing children = increased activity when listening to non-random syllables vs random syllables

autistic children = no increase

Question 24

what do brain activations indicate when listening to random / nonrandom syllables?

Answer 24

indicates differentiation of random and patterned sounds

more severe autism = less differentiation

Question 25

As brain activation decreases -

Answer 25

communicative difficulties increase

Question 26

autistics do not show a difference in brain activity when they listen to random and non-random syllables

Answer 26

this may help to explain their symptoms (the difficulty in detecting statistical regulalities)

Question 27

what are frequent frames?

Answer 27

gramatical categories that occur in predictable positions in a sentence

Question 28

what did Mintz say about frequent frames?

Answer 28

these frames occur reliably in child input = help children learn about grammar