Introduction: Verbal WM and sentence comprehension

1
Introduction Verbal WM and sentence
comprehension

• COGS 551
• Human Memory, Spring term 2007, METU
• Annette Hohenberger
• David Caplan and Gloria Waters (1999)?

2
Verbal Working Memory
Caplan Waters Is there a further
fractioniation of the verbal working memory for
sentence comprehension?

• Logie_Lecture_Notes
• psy-www-old.psy.ed.ac.uk/Local/Lectures/Psych3blog
  ie/Level3Lecture8.ppt -

3
2 types of verbal WM

• Interpretive processing
• ?parsing
• (lat pars piece)?
• Recognizing words, their meanings and syntactic
  properties
• building up syntactic and prosodic
  representations
• assigning thematic roles

• Postinterpretive processing
• Storage of verbal info in LT semantic M,
  reasoning, planning actions

Since parsing is an autonomous process, in the
view of modularists (Fodor, Frazier, Forster,
Garrett), the idea is to also grant it a separate
verbal WM component
Parsing is a reflex (Garrett)?
4
Sentence comprehension and WM

• Certain S structures are harder to parse than
  others, e.g. Center embedded S
• The man that the woman that the child hugged
  kissed laughed
• The child hugged the woman
• The woman kissed the man
• The man laughed
• All theories of parsing assume that complex
  sentences such as center-embedded S's require
  more WM resources.
• Parsing time is a function of S complexity.
• In experiments, duration of eye fixation,
  duration of self-paced word-by-word reading,
  lexical decision time increase when parsing
  complex Ss.

5
Interpretive processes require WM

• Increased demands on interpretive processes
• Subj-Rel-Clauses
• The boy that hugged the girlSubj-Rel kissed the
  baby
• Boy boy hugged girl kissed baby
• Obj-Rel-Clauses
• The boy that the girl huggedObj-Rel kissed the
  baby
• Boy girl hugged boy kissed the baby
• Obj-Rel-Clause is harder to parse. Why?

6
Post-interpretive processes require WM

• Increased demands on post-interpretive processes
• Clause with one proposition
• The boy hugged the girl and the baby
• Clause with two propositions
• The boy hugged the girl and kissed the baby
• The clause containing two pospositions is harder
  to process than the one with just one
  proposition. This increasing demand, however, is
  qualitatively different from the one between
  differently complex syntactic clauses (previous
  slide). It affects post-interpretive processes,
  e.g., storage in LTM.
• How are the two demands different?

7
The two hypotheses

• Single resource theory, SR
• There is just one WM for all verbal tasks
• Just Carpenter 1992 MacDonald et al. 1992,
  Miyake et al. 1994?

• Separate sentence interpretation resource, SSIR
• There is a special verbal WM component just for
  parsing
• Caplan and Waters

8
S comprehension and WM in normal subjects two
different methods

• Individual differences in WM span and parsing
• High-span and low-span subjects process simple
  vs. complex sentences

• Mutual interference between two tasks WM task
  and parsing
• (Unselected) Subjects are tested on two
  concurrent tasks, one requiring general verbal
  WM, one requiring interpretive parsing

9
Predictions of the two accounts

• 1. Method
• Individual differences in WM span and parsing
• High-span and low-span subjects process simple
  vs. complex sentences
• SR predicts an interaction both groups should
  differ only in the complex condition since it is
  there where the shortage of the common verbal WM
  resource becomes critical
• SSIR predicts no such interaction both groups
  should not differ since despite different verbal
  WM spans, there is a different resource pool just
  for parsing which is unaffected by WM span. So
  even if there is a shortage of verbal WM
  resources in low-span subjects, this does not
  affect their performance

10
Predictions of the two accounts

• 2. Method
• Mutual interference between two tasks WM task
  and parsing
• SR The concurrent verbal WM task should
  interfere with parsing, since both tasks draw on
  the same WM resource.
• SSIR The concurrent verbal WM task should NOT
  interfere with parsing.

11
No compelling evidence against SSIR from this
method

• Caplan and Waters (1999) doubt that there is
  compelling evidence for the Single Resource SR
  account.
• The fact that low-span subjects can remember less
  words in a self-paced reading-span task does not
  have to be due to weaker WM but may be due to
  less effective allocation of attention . In this
  case, the difference between the groups may still
  be compatible with SSIR

12
Combination of both methods

• Low- and high-span subjects perform 2 concurrent
  tasks 1 for verbal WM and one for parsing
  (simple vs. complex S)
• Predictions
• SR predicts a 3-way interaction Impairment of
  low span subjects on complex S's is exacerbated
  (espacially worse) by concurrent load.
• SSIR predicts no such 3-way interaction Low-span
  subjects do less well on the task overall, but
  not differentially worse on complex S's.

13

• ??The study of Fedoreko, Gibson, and Rohde (2006)
  tests a similar prediction
• ??Next presentation

14
References

• Caplan, D. and Waters, G. (1999) Verbal working
  memory and sentence comprehension. Behavioral and
  Brain Sciences 22, 77-126.
• Fedorenko, E., Gibson, E.,and Rohde, D. (2006)
  The nature of working memory capacity in sentence
  comprehension Evidence against domain-specific
  memory resources. Journal of Memory and Language,
  54, 541-553.
• Just, M.A. and Carpenter, P.A. (1992) A capacity
  theory of comprehension Individual differences
  in working memory. Psychological Review, 99,
  122-149.
• MacDonald, M. C., Just, M. A., Carpenter, P. A.
  (1992). Working memory constraints on the
  processing of syntactic ambiguity. Cognitive
  Psychology, 24, 56-98.
• Miyake, A., Carpenter, P., Just, M. (1994). A
  capacity approach to syntactic comprehension
  disorders Making normal adults perform like
  aphasic patients. Cognitive Neuropsychology, 11,
  671-717.