Visual processing of words
- Fixation: 200-250 ms (80% of content words)
- Saccades: 20-40 ms (9-7 letter jumps)
Assumptions of visual processing of words
- Immediacy: reader attempts to interpret each word as it is fixated
- Eye-mind: Eye remains fixated on word as long as it is still being processed
Gaze-contigent paradigm
- Cover foveated letters
- WPM dropped from 300 down to 50
- Shows that fixating on words is critical
Skipping words
- Usually a function word (the, and, but)
- Skip more often with easy text
Preprocessing words
- If change next word during a saccade then people don’t notice
- Shows that very little of the next word is preprocessed
Fixations and reading skills
- Fixation patterns differ between readers’ skill levels
- Lower skill = more fixations, backtracking, longer fixation time
How to model word recognition
- Center model around robust “marker” effects
- Start to build around those effects
Marker effect in word recognition
- Word frequency (common vs less common words)
- Word frequency affects fixation time, lexical decision time, and naming time
Logogen
- 1 logogen per known word
- Orthographic and phonological attributions of each word
- “Information” acculmiator
- Each have resting activation level
- Each has activation threshold
Explaining frequency effect
Logogen threshold lower for common words
Repitition effects
- Recognition faster the second time you see a word
- No repitition effect when reading house then seeing a house
Stimulus quality and word recognition
Degraded presentation results in longer retrieval time in lexicon decision time
Role for phonological coding during reading
- EMG (electromyography) during silent reading, lower for easy text
- Sound priming occurs, visual priming does not
- Pseudohomophones retrieval worse than non words for naming and better than non words for lexicon decision time
Why speech codes available during reading
- Keep info active in working memory
- Derive meaning in individual words
- Meaning of sentences
Lexical route of naming non words
- Sensitive to frequency
- Provides correct phonological codes for regular and irregular words
Nonlexical route of naming non words
- Not sensitive to frequency
- Correct code to regular words
- Incorrect code to irregular words
Learning to read: whole words vs phonics debate
- Set of logographic-looking words, containing hidden alphabetic system
- Trained to read aloud: whole word vs phonics group
- Initial trials showed whole-word group was better than phonics group
- Later trials showed whole-word group performance decreased and phonics group performance increased
