PSY2002 W2 Language II: Speech Perception - L Flashcards
Bottom up processing
Sensory input up to semantic understanding.
Top down processing
Already estabilished knowledge apply down to sensory input.
Mental lexicon access
You access your mental lexicon when you process speech. You access the phonological which will then access syntax, then access semantic meaning and that might also access orthographic meaning
Categoricla perception
Ability to distinguish between sounds on a continuum based on Voice Onset Times (VOT) (holding onto your throat) - Va vs. Fa
What’s word length’s influence on speed of lexical access?
long word, slower to process shorter words take less time to access (DOG/MICROSCOPE)
Perceptual learning
Adjust categorical percepiton based on scounds we hear.
What does frequency have to do with speed of lexical access?
the more frequently word is accessed in lexicon the quicker you can access it.
Example of cohort model
If a word starts with “AP” it will be activated [apple, apart, ape, April] as we get more information, we start to disactivate words that no longer match the acoustic information.
What type of processing happens (cohort model?
Bottom-up processing, from sensory input to semantic meaning. We activate the cohort based on the initial phoneme.
What s the cohort model?
Lexical activation of the ‘cohort’ that match the input. Gradual de-activation of items that fail to match the input. Lexical activation, uniqueness point.
What is a uniqueness point?
is when all the other words have been disactivated, we only have one word left standing, [Apric] you can have the uniqueness point before hearing the whole word.
What does the TRACE model predict?
predicts that features activate phonemes that activate words with a gradual increase in activation of words that match all features so that the word with the most activation wins (option 3) [gradually activate words]
What are two models of speech comprehension?
TRACE model and the Cohort model
What does the cohort model predict?
predicts that we access words in the lexicon via activation of all words sharing initial features and gradually de-activate words that stop matching the features (option 2) [last word standing]
When matching sounds to mental representations what do we do?
Actiate words that match the sounds at each point in the unfolding speech stream, activated all matching wrds and gradually deactivate words that no longer match, gradually activate matching words until one word has more activatin than others.
What are some lexical characteritics that effect speed of lexical access?
Neighbourhood density, freuqency, word length
What evidence is in favor of the cohort model?
Gating Experiments ( Warren & Marslen-Wilson)
What does sentence context affect?
Integration is affected by sentence context. Sentence context does not influence the process of lexical access. Lexical selection is based on activation of phonology and semantic information.
What are the three stages of word recogntion, cohort model?
Access, selection and integration
What is the integration stage of the cohort model?
Semantic and syntactic properties of the word are integrated and checked against the sentence
What is the access stage of the cohort model?
Acoustic-phonetic information is mapped onto lexical items
What is the selection stage of the cohort model?
Candidate words that mismatch the acoustic input are de-selected - candidate word is chosen
What is the architecture of the cohort model?
“Facilitatory signals are sent to words that match the speech input Inhibitory signals are sent to words that do not match the speech input.
Bottom-up processing has priority “
What does the the gating experiments suggest ?
Suggest that recognition of a word is gradual process that starts from word onset and continues until the end of the word. Candidate words that no longer fit the acoustic input are eliminated
What is the Gating experiments?
Participants are presented with fragments of words that gradually reveal the whole word and asked to guess what the word is after each presentation. Let’s try this out! Wooclap, you will be shown a fragment of a word, complete it! (keep in mind, we can see who has entered what).
How does the cohort model process context?
Showed the word nurse (target) then doctor (nurse) you are primed to activated fast. Doctor and Nurse are semantically related. Spreading activation allows ‘nurse’ to become active when ‘doctor’ is presented. Sheep and Nurse are not semantically related. Presentation of the word sheep does not activate the word nurse.
What is the cross model priming Zwitserlood, 1989?
Prime word auditory, target word visual. /captain/ - SHIP. Related prime target pair. Unrelated prime target pair – WICKET. Prime you with the word /cap/ that could be captain or capital related prime target pair – Money and ship. The impact of context – biassing – hearing the sentences.
How are words recognized in the TRACE model?
In TRACE words are recognized “incrementally by slowly ramping up the activation of the correct units at the phoneme and word levels”
What does the revised cohort model change from the original?
Context influences selection/integration fo word into sentence. The word with semantic activation that fits the context of the sentence will be integrated into the sentence.
What is the cohort model Predictions?
Items that match acoustic input but do not match sentence context are activated. Items that match acoustic input but do not match sentence context are deactivated once the word is selected.
Summary of the revised cohort model?
Speech perception is based on matching acoustic input to stored representations of words in the lexicon. Words are recognised via a competitive process that activates a word ‘cohort’. Cohort candidates do not actively engage with each other. Words are identified when they reach their uniqueness point. Cohort candidates that do not match acoustic input are eliminated. Context does not constrain activation of initial cohorts but allows for rapid elimination of candidates that do not match sentence context
What did Zwitserlood find evidence for?
Zwitserlood did find evidence for a faster reaction time for relate concept; (SHIP/MONEY) and slower reaction for unrelated concepts (WICKET).
What are three aspects of the TRACE model?
Gradual activation of item that matches the imput, lexical competitive inhibitoin, rhyming competitor.
What is the visual word paradigm?
The evidence is in agreement with the TRACE model. Participants looked at the beaker the beetle and the speaker. (Allopenna et al., (1998). Participants also looked at the speaker which shows evidence for the TRACE Model.
What is the activation of words in lexicon from allopenna et al?
Using an eye tracking study demonstrated that words with overlapping phonology that do not start with the same onset as the speech input (rhyme competitors) are activated in speech perception
What evidence is there for the TRACE Model?
Activatio fo words in the leixon Allopenna et al, 1998.
Are words activated based on shared word initial sounds?
The evidence from Allopenna et al. (1998) and others suggests that words that rhyme with sounds in any part of a word may become activated
What are the differences between the TRACE and the cohort model?
TRACE emphasis top-down processing, Cohort minimise the impact of top-down processing, Cohort predicts elxical access is biased towards activation of words with shared onset, TRACE accomodates the activate of rhyming competitors.
What does the TRACE model no account for?
How the context migth afect speech perception
How are words activated under the cohort model?
words are activated if they match the onset of incoming acoustic information.
What does the cohort model give priority to?
Bottom up processes
What is inhibited in the TRACE model?
Inhibitory signals inhibit irrelevant information within levels
How does Top-down processing interactive with the TRACE model?
Facilitatory connections between levels also travel down from the lexical level to the phoneme level and the feature level = Top down
What’s the TRACE model?
Implemented computaional model based on connectionist principles, processing nodes correspond to mental representations of features phonemes and words.
How does Bottom up processing interactive with the TRACE model?
Each level is connected via facilitatory connections. Activation spreads up from features to lexical items = Bottum up
What’s the influence of neighbourhood density to the speed of lexical acess?
Lots of neighbours, processed more slowly
In the Dell model of speech production based on spreading activation what do we mean by interactive?
Phonology and semantic processing occur in parallel.
Evidence of semantic and phonological competition in picture naming tasks which model of speech production best accounts for the finding?
Dell because of an overlap
Which models of speech production would be described as discrete?
WEAVER++ and Fromkin’s model of general utterance.
WEAVER ++ When items are processed to the LEMMA level, what happens to other items that are semantically related that may have been activated also?
They are inhibited from receiving further activation.
What are some challenges to lexical access?
Continuous speech stream (different patterns/speed) words flow into one another.
Homonums and Homophones
Co-articulation
Different accents
Invariance problem
What are Homonyms?
Bank ($) vs Bank (river)
What are homorphones?
Aisle vs Isle
What is co-articulation ?
speech production influenced by sounds that proceed and follow a phoneme (b in bill, ball and able - dexterity)
What are invariance problems?
Problems of definition of acoustic properties - phonemes, syllable, words
What is the cohort model ?
Marslen-Wilson 1987
predicts that we access words in the lexicon via activation of all words sharing initial features and gradually de-activate words that stop matching the features [last word standing]
What is the neighborhood effect?
words that match the acoustic input compete for activation. Neighbors compete with each other for recognition [apricot – aprikol]. The uniqueness point has change because of extra competition. Learning ‘aprikol’ slows down recognition of the word Apricot. Words with high frequency have high resting states - less activation required to recognize high frequency words. Apricot would be recognized more quickly than a low frequency word like aprikol, because of the frequency effect.
