Word Reading and Dyslexia Flashcards
The word identification problem
A typical reader can identify 2-3 words/second.
• Even at presentations of 10 words/second,
meanings are activated.
• Identification takes in the region of 200ms./word
• Once you’ve learnt to read, you can’t stop yourself
– E.g. the Stroop effect.
How do we recognise words?
- Direct access hypothesis: visual pattern
recognition.
• We can go straight from visual processing to meaning
without sounding out words, otherwise we would have
trouble with: - Indirect access/phonologically-mediated
hypothesis: translation from print to sound.
• we can also sound out letter strings, as we can
read non-words, like tegwop, burd. - Dual-route hypothesis: either of the above
routes.
How is word recognition assessed?
Naming aloud or lexical decision. lexical decision task, subjects have to
respond “yes” rapidly when the see a word, “no”
when they see a non-word.
What makes a word easy/hard to read?
- Spelling-to-sound Regularity.In general, irregular words take longer to read aloud (name)
- Frequency (how often a word occurs)
- Semantic Priming (context)
• Single Words
The Logogen Model
first incarnation, Morton, 1969
An example of a Parallel Search Model. Essential features:
• Every word we know has its own simple feature counter
called a logogen corresponding to it.
• Information sent in parallel to all the word detectors
(logogens) in the lexicon.
• The more information that is consistent with a word the
greater the activation of the logogen.
• A logogen accumulates evidence (more activated) until its
threshold level is reached, then the word is recognized.
• More frequent words have lower thresholds.
• Recently read words/associates have lower thresholds
Morton’s (1969) Logogen model
Logogen model: some issues:
• The simple logogen model could not explain
effects of ‘higher level’ information on ‘lower
level’ processing (classic example: sentence
context effects).
• The logogen model attempts to explain how we
recognise words, but has nothing to say about how
we read words or non-words aloud.
• Nonwords can sometimes be rejected very rapidly
How many routes to reading aloud?
Dual Route Model: driven by the idea that there are
two aspects to reading: rules and memory.
• One route needed for reading irregular words (e.g.
yacht, pint). The lexical route
• One route needed for reading non-words (e.g. blik).
The sub-lexical route.
Evidence for two routes
Pronunciation of nonwords cannot be done by the Lexical
route (no lexical entries for them)
• Pronunciation of exception words cannot be done by Sublexical
route (e.g. aisle)
• Two different variables: Word frequency affects the lexical
route; regularity (consistency) affects the sub-lexical route.
Neuropsychological Evidence for Dual
Routes: Dyslexia
Dyslexia: An impaired ability to read, which can’t
be attributed to more general factors such as visual
impairment, low intelligence, motivation.
• Acquired dyslexics: people who developed
normal reading ability but have lost the ability to
read properly, usually as a result of brain injury.
• Developmental dyslexics: people in whom the
ability to read has never properly developed.
The major acquired dyslexias
Surface dyslexia (e.g. Marshall & Newcombe, 1973, patient JC)
– Impaired ability to read irregular words (e.g.
have, pint, colonel)
– Regularisations (e.g. “kol-o-nell” for colonel)
• Phonological dyslexia
– Impaired nonword reading
– Relatively preserved ability to read words
Dual route theory and dyslexia
The dual-route model accounts for:
• Surface dyslexia by assuming that it is caused by
damage to the lexical route (the intact sub-lexical
route can deal with regular words and non-words).
• Phonological dyslexia by assuming that it
represents damage to the sub-lexical route (the
intact lexical route can deal with words).
