Flashcards
Word Recognition and Reading
Cards in this deck (51)
- 1Facts about reading*culturally acquired skill
* instruction required
* 10% adults Aus, UK, US functionally illiterate (read below grade 4) - 2Reading Highly automatic* 250 words per min (WPM) silently
*identify word in 100ms
*read aloud in 500-600ms
*judge word meaning in 600-900ms
* needed to function well in society - 3Reading complex skill* involves perceptual, cognitive, linguistic processes
- 4Linguistic processing skills for reading* orthogrphy- spelling patterns
*phonology- sound of words
* semantics_ meaning of words
* syntax - grammar - word order
*Higher level discourse integration- making inferences- bringing in world knowledge - 5Various tasks psychologists for different kinds of processing to study reading* lexical (vocabulary) decision tasks uses orthography ( is a letter string a word or not)
* semantic categorization > semantics ( Is this word an animal or a tool?)
* reading aloud > phonology ( need to translate print into sound) - 6Research Methods RT* RT dependent variable in
- lexical decision tasks,semantic categorization task and naming task (read aloud) - 7Research Methods Indirect measures (RT) Automaticity Stroop colour naming* Interfernce tasks > strop colour naming
* where subject names the colour but word green written in red
* to study how the word interferes - 8Research Methods Indirect measures (RT) Automaticity Priming* semantic priming, masked priming
* priming how what follows before influence the target e.g. car> truck how car influences truck
* masked - briefly flashed prime item that subject not aware of - 9Research methods* record eye movements
* neuropsychology> study of patients who became reading impaired from brain damage
* neuroimaging > looking at brain with machines while you read > ERP (EEG) MEG< fMRI - 10Automatic Processing FAST INVOLUNTARY UNCONSCIOUS* Fast reaction Times
* responses involunatry/ unavoidable
eg stroop tasks although looking only at the colour the word interfere e.g. shown red written in green and very hard to ignore the word red
*Unconscious as shown in masked tasks - 11MASKED PRIMING#### Forward mask (500ms)
pudge Prime (50ms) not seen it is so fast
JUDGE target ( as long as needed)
reaction time to target quicker if the prime is related. like quicker for pudge than slant
See video at 24 mins - 12WORD RECOGNITION silent reading THEORIESMcCelelland & Rumelhart's 1981
Interactive Activation Model
the basis of most models - 13Interactive Activation ModelHierarchy of levels
3 level dectors 1 st feature (- / o) , 2nd letter (A) , 3rd word (CAT)
Bottom - up excitation/ inhibition and top - down
WIthin level inhibition- letter a would inhibit O and N - 14Interactive Activation Model* explained word superiority effect
- 15word superiority effect Consistent finding to IAM* identifying a letter is easier if it is in a word context
* explained by a top-down feedback
the word level feeds to the letter level exciting and inhibiting - 16Slot Coding Finding inconsistent with IAMIAM adopts a slot coding assumption
ie assumes letter positions coded precisely
*tell difference between cat and sat saw and was
* model has to show how cat and act differ
*model doesn't capture the similarity between words with letters in wrong order CAT> ACT as dissimilar as CAT > BUT - 17Problem with slot coding assumption*jumbled letters in text can still be read
*cambridge email
Aoccdrnig to rscheearch at Cmabrigde... - 18Transposed- letter (TL) priming effect*according to slot coding assumption
jugde>judge and junpe> judge would have same reaction time
* by actually if the prime just has transposed letter judge reaction time is faster (decide if judge is a word)
*masked-priming task proves this - 19Automatic phonological activation silent doesn't require to say allowed*phonology is retrieved automatically in silent reading
- 20Strong phonology hypothesis (1998)*phonological representation is necessary product of processing print
* phonological processing is mandatory/ obligatory/involuntary in reading
* you do have to recode print into phonemes - 21Experiments to support Strong phonology hypothesis Homophone interference* Van Orden (1987)
*used homophones bred/bread: sight/site
*Homophone interference effect
TASK: Subjects made many errors when asked to make semantic categorisation
Is it a flower? ROWS
Subjects mistook ROWS for ROSE
This shows they were phonologically processing - 22Experiments to support Strong phonology hypothesis Masked Phonological PrimingRastle and Brysbaert (2006)
Meta analysis
Good example
Kinoshita & Norris (2012)
*word processing faster with phonological identical non words klip > clip than with with primes of similar orthography plip > clip
even when primes are so quick they are not consciously seen
Task: lexical decision Is it a real word?
Conclusion: phonological processing rapid and automatic - 23Reading Aloud* english alphabetic system
* letter or letter clusters (graphemes) smallest unit of sound: phoneme
* grapheme b, ee, sh, igh
Not pl (that is two sounds, two graphemes - 24Orthographic Depth* English is very deep or opaque because the spelling patterns are not regular
eg ough in though, cough, tough
* same spelling pattern, three different sounds
* irregular mapping Pint does not rhyme with mint, hint
* italian is very shallow/ transparent you sound it out as it is spelled. one cluster of letters has only one way to say it
one to one mapping - 25Orthographic depth and reading development* reading development is slower in deep orthographies
- 26Not all letter strings are read same way* regular words cat, fog
*irregular words comb, pint ( don't follow the sounding out
*pseudowords non words that can be pronounced e.g. junk - 27Reading words vs pseudo words (non words) Different procedures usedPINT vs JINK
* skilled readers can read both
*JINK read by mapping spelling to sound rules
*PINT violates mapping from grapheme to sound rules. Is irregular
Has to be stored in the lexicon (mental dictionary.. vocabulary)
need to use the lexical route to read word aloud - 28Dual- Route Cascade Model for reading aloud* Coltheart et al (2001) developed computational DRC Model
* ways of reading leads to the DRC model
*Route One non-lexical (sound it out by mapping graphemes to phonemes
* Route 2 and 3 lexical route - 29Two non lexical routes* route 2 print > orthographic analysis (letters) > lexicon> phonological system > response
*route 3 also goes via the semantic system so get the meaning of the word
* difference between route 2 and 3 not very clear
* lexical route sensitive to frequency - 30DRC predictsregularity x frequency effect
response time
EG
COME irregular = COLD regular
because COME has been seen so many times
low frequency words
irregular > Regular for RT
PINT > Pink - 31DyslexiaDRC to explain Dyslexia
Aquired Dyslexia - brain damage
Development - struggling to learn to read - 32Types of Acquired dyslexiaSURFACE
intact non word, poor irregular
apply regular rules
PHONOLOGICAL DYSLEXIA
intact word reading, poor at non words, can't sound out
DEEP DYSLEXIA
like phonological plus semantic errors
TULIP for ROSE errors relate only to meaning - 33DRC & DYSLEXIA neurological case study Surface DyslexiaMc CARTHY & WARRINGTON 1984
*non word reading accuracy 100%
*regular words 81%
*Irregular words 41% hatch, pint errors usually reading irregular word to rhyme with spelling pattern
Coltheart says:
USING only Route 1 the non lexical route to read
"knows" read as canows (canals) - 34DRC & Phonological dyselxia* can't not read non words and unfamiliar words
Patient WB (Funnell, 1983)
*cannot sound out letters
could read real words with 85% accuracy
*couldnt map graphemes to phonemes
* using only route 2 and 3
*meaning activated and sound is generated in the phonological output lexicon - 35DRC & DEEP DYSLEXIA DRC doesn't explain that well Coltheart 2001 deep dyslexia outside scope of DRCproblem reading unfamiliar words and non words with semantic errors
* ship > boat
* abstract words and function words particularly difficult (the>and)
* when recover often become phonological dyslexias - 36Connectionist Triangle Model* all information used to read real and non words
* othrography: phonology : semantics cooperative and competitive interactions
*two pathways
*Direct orthography > phonology
*Indirect through semantics
* Does not assume an orthographic lexicon (store of knowledge of spelling of words)
* Based on distributed knowledge
If one is impaired other areas become more involved - 37Connectionist Triangle Model Surface dyslexia*Woollams, Lambon-Ralph, Plaut & Patterson (2007)
* damage to semantic system
* strong association between impaired semantic knowledge and surface dyslexia
* if semantic knowledge intact can read irregular words - 38Connectionist Triangle Model Regularity vs consistencyRegularity (DRC)
*applies to words only
* regular words follow grapheme-phoneme mapping rules. regular or irregular
Consistency (Connectionist Model)
* applies to words and non words
*consistent word/non word: pronunciation agrees with similar words gaze similar to haze, maze
Tave inconsistent : similar to save, gave cave BUT "have" - 39Regularity or consistency? Connectionist say consistency more important* most irregular words also inconsistent
Jared (2002) showed consistency more important
BEAN: Regular and consistent
BEAK: Regular and inconsistent (steak)
*quicker to read words that are consistent in both high frequency and low frequency words - 40Eye Movement Studies Eye Tracking*4 words a sec
* whats happens when we read sequences of words: sentences
* Use eye trackers to
-watch how eyes move
-and control what you see
* can test
-whats happens when read sentence?
- how eyes move
-do we only process where we look? - 41Eye movements* saccaades
* fixations
*Regressions
*end of line sweep - 42Saccades* rapid eye movement
*25-30 secs jumps tend to move 7-8 characters
*you "see" nothing at all during a saccade (jump)
* ballistic- can't change direction till the saccade is finished
*separated by fixations - 43Fixations* 250ms
Dont fixate on every word
* tend to skip function words like "the"
high frequency short words
*effective field of view is 3-4 letters to left of fixation and up to 15 letters to the right in left to right reading - 44Regressions*infrequent right to left movements of 2-5 characters
(we read left to right) - 45End-of-line-sweep*single long saccade fromend of one line to beginning of next
- 46What to Fixate?* Common, short, predictable words not fixated (skipped)
*80% of content words (meaning): nouns, verbs, adjectives
* 20% of function words - articles (a, the)- conjunctions (and, but, or) - pronouns (he, she) - 47Applications of eye movement studies SPRITZ* March 2014 Spritz announced a reading technology to speed read at 500 wpm
* Claimed that 20% of time processes content and 80% is moving the eyes from word to word
* eliminate time scanning by presenting words centred around a Optimal recognition point (slight to left of centre of word
* eliminates saccades by using RSVP (Rapid Serial Visual Presentation) : one word at a time
THESE CLAIMS INCORRECT - 48SPRITZ CALIMS Optimal Recognition PointEvidence comes reading words in sentences
* no evidence in isolated words that fixations land slightly left of centre
* no evidence that saccade time reflects search for ORP - 49RSVP Reading rapid one word at a time* Comprehension and memory fail at RSVP rates of 500 wpm especially for material longer than sentences
* reading rate not constant - 50Reading Rate depends on* type of material
light fiction> newspaper> psychology, physics
* readers control rate within sentence as a function of comprehension difficulty
*Fixation longer for more difficult words: Low frequency and less predictable words - 51SPRITZ evaluation* little theoretical basis
* confuses preferred landing position in normal reading with Optimal Recognition Point
* Comprehension at rapid RSVP rates breaks down
-rate varies as function of difficulty
Information
- Description
- Flashcards on Word Recognition and Reading, created by Vicki Hanna on 14/05/2016.
- Created by
- Vicki Hanna
- Views
- 2
- Rating