Language and Cognition Flashcards
What are the differences between semantic and episodic memories
episodic = more prone to forgetting and brain injury
episodic = re-live to recall
semantic recall = automatic
what is the lexical decision task, and what are factors that affect the RT of the lexical decision task
lexical decision task = give a string of letters > then have the participants to identify if that is a word
factors affecting the RT of the lexical decision task
pronounceability of the word
the higher = the harder
semantic priming
if the word comes after a word that has similar semantic meaning = faster
repetition priming
if the word is repeated, the second time RT = quicker
What are the causes of priming effects
spreading activation theory
that when the word is produced > activating the word, and other words that are associated = semantic priming
the activation and decay is quick, therefore the priming decays quickly as well.
word frequency effect
Explain what is the frequency effect (and the effects on lexical decision, recall and recognition)
word frequency effect = the more common the word is = the faster the RT is, the less common the word is = the slower the RT
Lexical decision task
HF words = quicker RT; LF words = slower RT
supported by eye-tracking data
recall task
effects only show on pure list, but not mixed list
better recall for HF words, worse recall for LF words
in mixed list the difference is mediated
recognition task
LF words show advantage (higher hit rate and lower false alarm rate) than HF
mirror effect
Explain what is the main flaw for the word frequency effect
there is not yet a set explanation for the word frequency effect
therefore doubting the credibility of the word frequency effect
the reason why there are so many explanations:
HF and LF words are different in so many ways
length, concreteness, and neighbour size (associate size)
What is the context variability theory, the explanation of it, and if the theory predicts episodic memory tasks (and other theories that the theory can explain)
states that not word frequency that predicts performance, but the variability of the word is used across different context
the more context = the quicker the RT for lexical decision tasks
Explanation: Rational analysis
that cognition and memory is shaped according to the need of the person in the environment
words that have a high CV = more likely they are used in the future
the theory also predicts low CV words have more advantage in recognition and recalling tasks
also explains primacy, spaced, recency, and words highlighted in different colours are more likely to be recalled
Explain what is the traditional theory on word identification (flaws) and what is the alternative proposal
emphasis on rules and exceptions
flaws
cannot explain when the rules are learnt
cannot explain the prioritisation of the rules
brain damage/ageing don’t take away rules (just graceful degradation)
cannot explain why context affects perception (THE CAT)
interactive activation model
computational model
three levels: features, letters, words
lateral inhibition: weakens candidates with less activation so that we only perceive one word/letter
interaction: feature > letter > word (bottom-up) > letter (top-down)
what is the TRACE model (what does it explain, right context effect, word segmentation, flaw)
explains speech perception (how to segregate words in speech without pauses)
phoneme level replacing letter
right context effect
sometimes when speech people don’t pronounce certain syllables, but still able to pick up the entire word
done by using the rest phoneme to guess/compensate the lost phoneme
word segmentation
the first word is pronounced
the second word is perceived when the information is inconsistent
then lateral inhibition and the first word is decayed
flaw
cannot explain semantic processing of speech perception
What are the first two proposals on acquiring language
behaviourism
learn through reinforcements
when we produce a sentence > reinforced > learnt
syntactic approach (Chomsky)
criticised the behaviourism approach
poverty of stimulus
we produce sentences that we’ve never heard before
universal grammar = we map languages onto the map > further production
syntax = emphasis on the importance of sentence structure
we use sentence structure to understand the sentence
flawed
sometimes is the other way round (i.e., the semantics of the sentence interprets how we interpret the sentence)
proposed syntax rules to be innate but no evidence
What is the parallel distributed processing accounts (the four principles, the framework when learning past tense, criticisms)
four principles
learn through error compared to the prediction
scattered around and with many connections (explained graceful degradation)
no new unit is added
the model doesn’t start with any knowledge
learning past-tense
first the present tense is encoded > converted into wicklefeatures > produced > if error = make corrections
criticisms
too hard to understand
sensitive to training sets
does not success on all words
learn things that people can’t learn
does not solve the double dissociation problem
