Models of Word Recognition and Production Flashcards
T or F? there are no gaps between spoken words?
T
T or F? the same spoken word will sound the same across different contexts
F
T or F? As you hear the sounds of words unfold over time, your brain is trying to predict what the word is before it ends
T
what is a cohort of words?
a possible number of words that start with the same sound
what is the problem of parallel activation?
the brain generating a number of different hypotheses in parallel of what the word is, after it hears the start of the word.. words become activated in parallel
what is a lexicon?
a part of the brain that is devoted to representing your knowledge of words, a metaphorical mental dictionary
what are the 3 linguistic modalities?
phonology, orthography and visuo-spatial/gestural
what is the first/most fundamental part of language processing?
recognising the forms of words, either their spoken or written form
what are lexical representations? (“lemmas”) stored?
our lexicons
perceptual forms of words map onto what?
our lexical representations
what do lexical representations do/represent?
Examples?
what underlies the form of the word
eg, its role, the way it can be used in a sentence, concepts (semantics) it maps onto
how long do spoken words approx take to unfold?
500 milliseconds
what two tests can provide some evidence for parallel activation of lexical competitors?
eye tracking and mouse tracking
what does cross modal priming refer to?
how spoken input can affect the way we interact with and are primed for different written input (written words)
what does cross modal priming measure? and how?
it measures which words are activated in the lexicon during the spread of activation and measures this by response time to recognition of a word fragments that do or do not prime associated words
who introduced the TRACE model and when?
McClelland and Elman, 1986
T or F. TRACE model is a connectionist model?
T
what are the 3 levels of the TRACE model of spoken word recognition? from lowest to highest
features (eg voicing of a sound), phonemes, word layer/lexical level
what can the TRACE model not explain
how meaning is attached to words, only focused on recognising individual words
why is the TRACE model considered an interactive activation connectionist model?
because there is activation between all layers of the model that influence and feed back on the nodes in other layers… there is both bottom up and top down functioning
what does threshold refer to when talking about nodes in the TRACE model?
this threshold determines how much input the node needs in order to send activation to other nodes
what does resting level of activation refer to when talking about nodes in the TRACE model?
reflects the frequency that the node is activated.
what determines resting level of activation?
this resting level of activation is determined by the frequency that the nodes are used… nodes that are used a lot have a higher resting level of activation as it does not take much to push them beyond the threshold
what is “lexical feedback” within the TRACE model?
top down interaction between the lexical level and the phoneme level, which reinforces the phoneme level, thus reinforcing the likelihood of both nodes from each layer occurring
what are lateral inhibitory connections at the word level in the TRACE model?
connections within the model where the activation of one word node and suppresses the activation of another competing word node… “lateral competition”
what are the 2 key questions in regards to bilingual lang processing?
- is there independent lexica or an integrated lexicon?
2. is processing non selective or language selective?
what does non selective v language selective refer to?
non selective means bilinguals cannot choose to only operate in one language based on the context they are in, language selective theorises that they can choose to operate in (activate) one lang
based on the two questions of bilingual lang processing, what are the 4 possibilities of this phenomena in bilinguals? which one is it?
- selective, independent
- selective, integrated
- non-selective, independent
- non-selective, integrated (THIS ONE)
T or F, different brain areas are used to represent different languages in bilinguals
False
T or F, bilinguals represent lang in the same areas of the brain that monolinguals do
True
what did van Heuven et al (1998) investigate? How?
whether or not there is evidence for non selective processing of language in bilinguals.
-They had dutch-english bilinguals decide if a word was or was not an english word, but for some participants, neighbouring words were thrown in the mix to impact RT
what did van Heuven et al (1998) find?
that for bilinguals, the more dutch word neighbours the english words had, the slower the lexical decision time
what are the 3 levels of an interactive activation (connectionist) model for orthography? bottom to top
Feature, letter, word, Language nodes/language tags
what are examples of the feature level in the orthography model?
pen strokes, eg horizontal, vertical, dots etc
what are language nodes/tags in the orthography model?
nodes that accrue info about which language the task requires the bilingual to use, thus sending inhibitory signals towards representations of the words that are non consistent with target language
