PSY 357 Exam 3 || Graded A+.
Regarding the Ferreira & Clifton (1986) experiment: Be able to identify or describe the
following elements:
1. Theories/hypotheses tested: correct answers 1. Does manipulation of "editor" context in prior
paragraph change syntactic interpretations? Can context eliminate garden paths?
2. Participants read different paragraph contexts to introduce a context manipulation
3. IV-paragraph context (1 or 2 editors) DV-first-pass reading time at "agreed"
3
2. Task/stimuli used: correct answers Participants read different paragraph contexts to introduce a
context manipulation
3. Independent/dependent variables used in design: correct answers IV-paragraph context (1 or 2
editors) DV-first-pass reading time at "agreed"
4. Predictions made by different theories: correct answers Interactionism: immediate effects of
context on reading times at "agreed"/ If reader are introduced to 2 editors they will NOT make
minimal attachment guess but if given one they get garden-pathed (2 editor = fast/ 1 editor
=slow)
Syntax-first: no effects of context on reading times at "agreed" (2 and 1 editor context = slow)
5. Results of experiment:
Implications for Modularity: correct answers Support the modular, Garden-path model, no
immediate effect of context found at disambiguating region (agreed)
Which specific piece of evidence (i.e., specific finding from a specific language or set of
languages) suggests that Late Closure is not universal? How? correct answers Spanish, Dutch
relative clauses attached to NP1
, According to Frazier's GP model (1987), speakers of all languages should prefer NP2
interpretation
Make relative clause part of current phrase (actress)
Most English speakers put relative clause with NP@
However it turns out that many languages show an N1 preference (Spanish, Dutch etc.) so late
closure must not be universal
Trueswell et al. (1994)—Understand its methods, results, and implications. correct answers
Research question: Can semantics immediately resolve syntactic ambiguities?
Methods: suspected the semantic constraints in Ferreira and Clifton weren't strong enough so
they replicated the study with stronger words (MV/RR)
Results: Readers slow down more at "by the lawyer" with "defendant" takes more time to
process
Implications: So semantics of "evidence"must reduce syntactic ambiguity immediately
Just and Carpenter (1992) found that individuals with more working memory capacity
interpreted sentences differently than individuals with low WM capacity. Be able to identify what
that difference is correct answers Individuals with more WM are better able to immediately use
semantic information (evidence can't examine anything) to disambiguate toward the RR clause
interpretation (The evidence that was examined)
WM helps us integrate more pieces of information
also be able to identify a description of what role working memory might play in parsing,
according to Just and Carpenter (1992), as well as its implications for the Modularity Debate.
correct answers It's possible that parsing sometimes resembles a modular process when WM is
low and sometimes resembles an interactionist process when WM is high
Perhaps we should move beyond the modularity debate and focus on other aspects of the
psychology of sentence comprehension
Constraint-based and garden-path (i.e., "traditional") theories differ in many ways, but there is
one assumption both types of models make with respect to the final product (representations) of
Regarding the Ferreira & Clifton (1986) experiment: Be able to identify or describe the
following elements:
1. Theories/hypotheses tested: correct answers 1. Does manipulation of "editor" context in prior
paragraph change syntactic interpretations? Can context eliminate garden paths?
2. Participants read different paragraph contexts to introduce a context manipulation
3. IV-paragraph context (1 or 2 editors) DV-first-pass reading time at "agreed"
3
2. Task/stimuli used: correct answers Participants read different paragraph contexts to introduce a
context manipulation
3. Independent/dependent variables used in design: correct answers IV-paragraph context (1 or 2
editors) DV-first-pass reading time at "agreed"
4. Predictions made by different theories: correct answers Interactionism: immediate effects of
context on reading times at "agreed"/ If reader are introduced to 2 editors they will NOT make
minimal attachment guess but if given one they get garden-pathed (2 editor = fast/ 1 editor
=slow)
Syntax-first: no effects of context on reading times at "agreed" (2 and 1 editor context = slow)
5. Results of experiment:
Implications for Modularity: correct answers Support the modular, Garden-path model, no
immediate effect of context found at disambiguating region (agreed)
Which specific piece of evidence (i.e., specific finding from a specific language or set of
languages) suggests that Late Closure is not universal? How? correct answers Spanish, Dutch
relative clauses attached to NP1
, According to Frazier's GP model (1987), speakers of all languages should prefer NP2
interpretation
Make relative clause part of current phrase (actress)
Most English speakers put relative clause with NP@
However it turns out that many languages show an N1 preference (Spanish, Dutch etc.) so late
closure must not be universal
Trueswell et al. (1994)—Understand its methods, results, and implications. correct answers
Research question: Can semantics immediately resolve syntactic ambiguities?
Methods: suspected the semantic constraints in Ferreira and Clifton weren't strong enough so
they replicated the study with stronger words (MV/RR)
Results: Readers slow down more at "by the lawyer" with "defendant" takes more time to
process
Implications: So semantics of "evidence"must reduce syntactic ambiguity immediately
Just and Carpenter (1992) found that individuals with more working memory capacity
interpreted sentences differently than individuals with low WM capacity. Be able to identify what
that difference is correct answers Individuals with more WM are better able to immediately use
semantic information (evidence can't examine anything) to disambiguate toward the RR clause
interpretation (The evidence that was examined)
WM helps us integrate more pieces of information
also be able to identify a description of what role working memory might play in parsing,
according to Just and Carpenter (1992), as well as its implications for the Modularity Debate.
correct answers It's possible that parsing sometimes resembles a modular process when WM is
low and sometimes resembles an interactionist process when WM is high
Perhaps we should move beyond the modularity debate and focus on other aspects of the
psychology of sentence comprehension
Constraint-based and garden-path (i.e., "traditional") theories differ in many ways, but there is
one assumption both types of models make with respect to the final product (representations) of
