SLAT6827/LING2020 Linguistics

1
SLAT6827/LING2020 Linguistics Literacy L2
Literacy Semester 2, 2003

• Week 3
• The psycholinguistics of reading
• the view from the bottom

2
Reading as skilled information processingUnderwoo
d Batt (1996), Chapter 1.

• Identifying the processes that enable us to see
  word meanings in printed and written symbols and
  integrate these meanings into the ideas presented
  in sentences
• Reading is a complex skill comprised of component
  subskills
• Reading can be examined as a problem-solving
  activity
• Laboratory approach to reading

3
SIMPLE MODEL OF READING

• Letter perception
• Word recognition
• Word integration (idea formation)
• Construction of a mental model

4
The laboratory approach to studying reading.

• controls the presentation conditions so that each
  of the subskills can be studied in the absence of
  the influence of other subskills.
• observes the effect of known changes in the
  materials that are to be read, e.g., frequency.

5
Information and knowledge

• Information is contained in the stimulus and is
  available for us for gathering, while knowledge
  is information that has been gathered over time
  and can guide the collection of future
  information.

6
Types of knowledge affecting the collection of
information in reading

• Shape of orthography
• Possible letter combinations
• Frequency of letter combinations
• Immediate context
• Knowledge of the world

7
Reading has been described as a guessing game

• The captain told the mate to drop
• the an

8

• ...anchovies?

9

• ...anchor?

10
The processing of information

• A process is a cognitive action that transforms
  information from one form to another.
• Theories of pattern recognition
• Template matching
• Feature detection

11
How do we recognise and distinguish letters?

• Theories of pattern recognition
• Template matching
• Feature detection

12
Template matching

• Patterns are recognized on the basis of
  similarity to stored patterns.
• K K K k k k K k

13

• This is upside down but you can still read it.

14

• Feature detection
• Patterns are composed of features. Letters
  consist
• of combinations of features.
• Physiological evidence for signal detectors

15
Feature detection

• Visual search task (Neisser, 1964)
• Task Find target letter Z
• Z is easier to detect when embedded in string of
  dissimilar letters
• O Z C B
• than in a string of similar letters
• X Y Z M
• Due to activation of letters with similar
  features, here vertical and slanted lines.

16
Serial or parallel processing?

• Word superiority effect is strong evidence that
  the letters that make up a word are processed in
  parallel.
• Letters in words are often identified faster than
  letters presented alone.
• G easier to discriminate from D when
  presented in a word, e.g. WAGE, than alone.

17
Accessing the mental lexicon Phonological access
model

• Written word
• Visual analysis
• Letter-to-sound conversion
• Auditory Mental Lexicon
• Output

18
Phonological access

• Strings of letters must be recoded as words
• Evidence for recoding from visual to speech-based
  code.
• Pseudohomophone effect is evidence for activation
  of phonological form brain
  gtgtgt brane
• bird gtgtgt burd
• Pseudohomophones have longer decision times than
  other strings that do not form words, suggesting
  we need to transform each visual pattern into its
  phonological form.
• Caveat simple phonological route would result in
  pseudohomophones being accepted as words. Thus a
  spelling check is also needed.

19
Spelling-to-sound correspondence rules

• Grapheme-phoneme mappings
• Regular mappings, e.g., Japanese, Finnish
• Irregular mappings, e.g. English

20
Regular grapheme-phoneme mappings

• Languages like Japanese
• ???? marubatsu 0s and Xs
• ???? daikon radish
• ?? inu dog

21
Irregular grapheme-phoneme mappings

• Languages like English
• Same grapheme but different phonemes HAVE and
  GAVE
• Different graphemes but same phonemes BOAT and
  HOE WISHED and MISHAP
• BEAUTY, CLEAR, CREATE, DEAL, GREAT, HEARD,
  LIKEABLE, REACT, SÉANCE

22
Direct access via graphemes

• Written word
• Visual analysis
• Visual Mental Lexicon
• Output

23
Direct access via graphemes

• Irregularity of phoneme-grapheme mapping suggests
  that direct access to the mental lexicon via
  graphemes is possible
• Words that are spelled and pronounced the same
  but have different meanings LIGHT, SAW, MINT
• gtgt Pose difficulty for both phonological route
  and direct access models.
• Words that are spelled the same, but have
  different pronunciation and meanings READ,
  INVALID, TEAR
• gtgt Can be handled by direct access model which
  allows the recognition of meaning prior to the
  assignment of pronunciation.

24
Experimental evidence for direct access

• Word decision shadowing task (Kleiman, 1975)
• 
  Decision time
• Type of
  No shadow Shadow
• Decision True responses False responses
  (msec) (msec)
• Graphemic HEARD BEARD GRACE PRICE
  970 1,095
• Phonemic TICKLE PICKLE LEMON DEMON
  1,137 1,509
• Synonym MOURN GRIEVE DEPART COUPLE 1,118
  1,238
• gtgt Shadowing most disruptive for Phonemic task,
  much less disruptive for Synonym task.

25
Phonological coding

• Phonological coding may not be necessary to
  lexical access, but it is common.
• The effects of phonological codes are widely
  evident
• Priming.
• Peripheral processing.
• Eye movement.

26
Priming

• The identification of a target word is enhanced
  (primed) by a preceding homophone
• e.g. SIGHED primes SIDE

27
Peripheral processing

• Words printed to one side of a target word
  inhibit lexical decisions when the spoken form is
  related to the target.
• Target Distractor word
• PAIN VANE
• Note that the peripheral word is not reportable

28
Eye movements

• Contingent display studies allow words to be
  changed immediately prior to reading. Target
  words that relate to the replacement words are
  manipulated. Shorter eye fixations occur when the
  prior word was a homophone.
• The generous man gave every sent to charity
• sent replaced e.g., with cent or rack. When cent
  appears the fixation duration is shorter.

29
Regularising sound-spelling correspondences in
English

• More symbols would be required to match the
  number of phonemes
• Homographs would increase (FOR/FOUR/FORE)
• Morphological information would be lost
  (INHIBIT/INHIBITION)
• Word recognition would be more difficult without
  redundancies in spelling (EIGHT/ATE)

30
Skilled information processing and attention

• Processes that are automatic do not require
  attention.
• Processes that require attention are more
  resource demanding.

31
Processes that require attention are more
resource demanding.

• Identifying propositional structure of text.
• Development of mental model of text.
• Concurrent tasks make it difficult to decide if
  sentences like the following are acceptable
• PIZZAS HAVE BEEN EATING JERRY

32
Processes that are automatic do not require as
much attention.

• Practiced letter recognition
• Practiced word recognition
• Evidence from Stroop effect for automatic word
  recognition
• SMALL LARGE
• SMALL LARGE
• Automatic lexical access (Swinney, 1979)

33
Nassaji, H. Geva, E. (1999). The contribution
of phonological and orthographic processing
skills to adult ESL reading Evidence from native
speakers of Farsi. Applied Psycholinguistics 20,
2 , 241

• Dependent measures
• Reading comprehension
• Silent reading rate
• Single word recognition
• Independent measures
• Phonological processing skill
• Orthographic processing skill
• Syntactic processing skill
• Semantic processing skill
• Working memory
• Rapid automatization naming task (RAN)

34
Nassaji Geva (1999) Results

• Efficient lower level phonological and,
  particularly, orthographic processing skills, as
  well as higher level semantic and syntactic skill
  contribute to ESL reading.
• Efficiency in lower level processes make a unique
  contribution to ESL reading comprehension.
• The complex nature of reading connected text
  places a premium on lower level visual and word
  identification processes.)

35
Last slide SLAT6827 week3