Lecture 11: Comprehension and Written Language

Question 1

What does visual processing involve?

Answer 1

• Involves Eye movements (Gaze)
• Series of pauses (fixations) and jumps (saccades)
• -fixations ~ 200 - 250 ms (80% of content words)
• • saccades ~ 20 - 40 ms (7 – 9 letter jumps)

Question 2

Why do we engage in series of fixations and Saccades during reading?

Answer 2

We cant see anything when our eyes are moving, cannot process visual information just with saccades (saccadic suppression). Visual acuity best at foveated area (we have to fixate in order to process letters and words)

Question 3

What does saccadic suppression mean?

Answer 3

no information processed during saccades

Question 4

What are the two assumptions of visual processing?

Answer 4

1. Immediacy: reader attempts to interpret/process fully, each (content) word as they are fixated on it (amount of time a person is fixated on a word reflects the amount of time it takes to process it, they don’t save it for later or process earlier information instead, it is immediate. Differences in fixation time depending on difficulty of word)
2. Eye-Mind: eye will remain fixated on a word as long as the word is being processed

Question 5

What 4 pieces of evidence are there for the necessity of engaging in a series of fixations and saccades?

Answer 5

1) Gaze-contingent paradigm (Rayner, 1998)
- What happens if we cover the foveated letters such that only the parafoveal can be seen?
- reading rate drops from 300 wpm to 50 wpm
- shows that fixating words is critical
2) Skipping Words (Carpenter & Just, 1983)
- when skip a word, it is usually a function word (the, and, but)
- skip more often with easy text
3) During Normal Reading
- If change next word during saccade?
- people do not notice
- shows, that very little of next word is preprocessed
4) Fixations & Reading Skill
- fixation patterns of high skill vs. low skill readers differ
- low skill - more fixations
- backtracking
- longer fixations per word

Question 6

What is visual word recognition?

Answer 6

Encoding of words and activation of their corresponding orthographic, phonologic, and semantic representations

Question 7

What 3 types of tasks are used to examine word recognition?

Answer 7

• eye tracking
• Lexical decision (deciding if it’s a word or not)
• naming (how long does it take someone to name a word?)

Question 8

How do we build models for word recognition?

Answer 8

• center model around robust “marker” effects (variables or factors that have a strong reliable effect on what it is you are trying to look at)
• then start to build around these effects

Question 9

What are the marker effects in word recognition?

Answer 9

• word frequency
• • common vs. less common words
• word frequency affects:
• • fixation time during reading
• • lexical decision time (LDT)
• • naming time

Question 10

What is the structure of morton’s logogen model?

Answer 10

• Encoding –> Logogen System –> Output

Question 11

What is a logogen?

Answer 11

• 1 logogen per known word (Logo = word, Gen= genesis as in origin)
• includes orthographic (spelling) & phonological (sound) attributes of each word
• “information” accumulator
• each logogen has a resting activation level
• each logogen has an activation threshold (once you exceed that activation threshold, it fires and because of that we now know that H O U S E is house (i.e., the threshold for common words is low, we don’t need a lot of activation to recognize that word. Takes more time to reach threshold for uncommon words. Exposure changes thresholds)

Question 12

How does a logogen explain the frequency effect?

Answer 12

• assume threshold is lower for common words

Question 13

What are context/priming effects and how did they modify Morton’s model?

Answer 13

• Doctor- nurse vs. tree-nurse
• Morgan’s original model did not deal with semantic priming/account for meaning so he modified his model to include context/semantics
• • When you see a word like doctor, the context system accounts for similar words, nurse is semantically closely related so nurse also starts to fire (pre activation)

Question 14

What are repetition effects?

Answer 14

Visual/visual presentation:
- House-house (present a word twice in a short amount of time)
- Recognition faster 2nd time (because of repetition priming)
- Evidenced by a higher resting activation level
Auditory/visual presentation
- Spoken ‘house’ vs. visual ‘house’
- Logogen should account for this priming because he thought that logogens were visual and auditory. However, he did not find repetition effects for visual to auditory priming.
- Thus, He changed his model again to account for this (broke it into two separate parts: phonological lexicon and orthographic lexicon)

Question 15

How did repetition effects influence Morton’s Model?

Answer 15

• Encoding –> - Split the logogen system into two parts: orthographic and phonological
  Orthographic lexicon/ Phonological lexicon –>
• Output
• Context system still there

Question 16

What is the relationship between stimulus quality and word recognition time?

Answer 16

• Stimulus quality and word recognition
• • Degraded presentation results in longer RT’s in LDT and naming tasks
• Stimulus quality is additive with frequency (didn’t matter whether it was a high frequency word or low frequency word. Stimulus quality added time to the encoding stage)
• Can use additive factors logic to determine locus of stimulus quality effects

Question 17

What is the role of phonological (speech) coding during reading?

Answer 17

1. Edfelt (1960)
   - EMG (electromyography) during “silent” reading
   - EMG (easy text) < EMG (hard text)
2. Meyer, Schvaneveldt, & Ruddy (1974)
   - sound priming (bribe - tribe) occurs
   - visual priming (couch - touch) does not occur
3. Rubenstein, Lewis, & Rubenstein (1971)
   - pseudohomophones (brane, froot, phocks)
   - nonwords (brate, freet, snocks)
   Naming: RT(brane) < RT(brate) (RT is longer when naming/reading nonwords like brate)
   LDT: RT(brane) > RT(brate) (RT is longer when deciding if pseudohomophones are words because they sound like real words)

Question 18

Why are speech codes (i.e., the phonology of words) available during reading?

Answer 18

• keep info active in WM (AVL is the dominant code in working memory. When we read sometimes we need to go back to figure out meaning of words/sentences so we keep words active.)
• derive meaning of individual words (ambiguity)
• meaning of sentences

Question 19

How do we name nonwords and how has the model been adjusted?

Answer 19

• Morton’s model is a model of our mental lexicons, words we already know. So how do we account for nonwords because we can name them? Added another route to the model
• The lexical route, the main route, has stored orthography’s stored knowledge of full words and word representation
• The non lexical or sub lexical: it is called this because it does not deal with words, the units within it are at the subword level. It deals with visual stuff but graphemic parts of a word. We can break words down into single letter or multi letter graphemes. Maps graphemes onto corresponding phonemes (grapheme to phoneme conversion route, or assembly route). Assembles into an output. This route is based on regularity

Question 20

How does word regularity and word frequency interact?

Answer 20

• Regular (mint, hint, tint)
• Irregular (pint- irregular spelling to sound mapping)
• Two factors: frequency, regularity (is there interaction or additive?)
• We can see that Hf words are named more quickly than low frequency words (main effect)
• Overall irregular words are named more slowly than regular words (main effect)
• However, we also have an interaction between word frequency and regularity. For high frequency words it doesn’t matter whether the word is regular or irregular, we name them quickly. Low frequency irregular words (e.g., pint) are named much slower than a non frequency regular word (e.g., hint)
• Non lexical route is not affected by word frequency because it doesn’t deal with words, deals with graphemes and phonemes (this is a slow process, identifies graphemes, process them and convert them). Whereas, the lexical route (faster process) is affected by word frequency, quickly when it’s a high frequency word and slowly when its low frequency. When its low frequency and irregular the lexical and non lexical route take about the same amount of time.

Question 21

What is the lexical route?

Answer 21

• sensitive to frequency (high frequency words are named quickly whether the word is regular or irregular)
• provides the correct phonological codes for regular and irregular words

Question 22

What is the nonlexical (GPC) route?

Answer 22

• not sensitive to frequency
• correct code to regular words
• incorrect code to irregular words

Question 23

How has this research provided evidence for the whole world vs. phonics debate?

Answer 23

McCandliss & Schneider (1997)

• set of logographic-looking words, containing hidden alphabetic system
• trained to read aloud: whole-word vs. phonics groups
• Results
• • Initial trials: Whole word group better than the phonics group
• • Latter trials: Whole word group performance decreased and Phonics group performance increased