3.1 - Sentence Comprehension Flashcards
Is parsing immediate or delayed?
Immediate
It happens very quickly & unconciously
What is Parsing?
Determining the linguistic relationship among parts of a sentence
What is Just & Carpenter’s (1980) Immediacy Principle?
Immediately after a word is heard or read…
- The meaning is accessed from memory - It is placed into the assumed syntactic structure
If we parse a sententence incorrectly, do we have to go back and reparse it?
Yes
What is the Wait & See Approach to Parsing?
We wait until we have all necessary information then we interpret
It’s an error free approach but is time consuming
What are garden path sentences?
Sentences where part way through we realized we need to reinterpret them
“The horse raced past the barn fell.”
Garden Path Sentences support the _____.
Immediacy Principle
What is the Modular Approach to parsing?
There are three modules: Phonology, Syntax, Semantics
Comprehension is the result of modules
In parsing, only syntax is considered. Semantics is not considered
What is Late Closure?
We attach the next piece of syntax to the phrase currently being processed
We “Close” clause boundaries at the last possible moment
Example: “Since Jay always jogs a mile seems like a very short distance to him.”
“Because Jay always jogs a mile…” <– We should have closed
What did Frazier & Rayner do in their 1982 eye-tracking experiment (Jay jogs)?
Used eye-tracking
Gave people the first parts of sentences: “Since Jay always jogs a mile…”
Then gave them either
- “…seems like a short distance to him.”
Late Closure fails!
- “…this seems like a short distance to him”
Late Closure works!
What did Frazier & Rayner discover in their 1982 eye-tracking experiment (Jay jogs)? What did this support?
(2)
Found that people had longer fixations in the first phrase ending (the garden path, Jay jogs)
This supports late closure
What is the Minimal Attachment (MA) theory?
That we attach items in a way that creates the fewest number of new constituents
Example:
“Tracy picked up John and his friend…”
We expect a NP conjunction (“…at the bar.”) but we can also end this with a new NP (“…went back to the bar”)
We prefer the 1st type of ending (NP conjunction)
Does the Minimal Attachment (MA) theory state that sentences are harder to process when there are more constituents?
Yes
What did Frazier & Rayner do in their 1982 sentence comprehension experiment (spy-cop)? What took subjects longer to read?
Gave subjects the sentence
“The spy saw the cop…”
Then gave one of two possible endings
“…with the binoculars but the cop didn’t see him.”
“…with the revolver but the cop didn’t see him.”
The binoculars sentence took subjects longer to read
What is the Interactive Approach to parsing?
That Parsing involves more than just syntax
What is MacDonald et al’s (1994) Constraint-Based Model?
We use all available information when we parse sentences (Syntactical, lexical, contexual, etc.)
Syntax & semantic interact during comprehension
What did Rayner et al do in their 1983 experiment (flowers)?
Gave people less plausible sentences (“The florist sent the flowers was very pleased”) and more plausible sentences (“The actor sent the flowers was very pleased”)
What did Rayner et al find in their 1983 experiment (flowers)? What theory did this support?
People experienced garden paths on both florist sentences. Plausibility didn’t matter.
Supports modular theory
What did Trueswell et al do in their 1994 eye-tracking experiment (evidence)?
Gave subjects two sentence beginnings
- The defendant examined…
- The evidence examined…
Then gave the same ending
- …by the lawyer turned out to be unreliable.
What did Trueswell et al learn in their 1994 eye-tracking experiment (evidence)? What theory does this support?
Subjects had an easier time with the sentence beginning with defendant than the evidence one.
Supports the interactive model
Do we use our semantic knowledge to determine if Minimal Aattachement is possible?
Yes
What did Ferreira & Clifton do in their 1986 experiment (editor)?
Gave subjects one of two sentences
- The editor played the tape and agreed it was a big story
- The editor played the tape agreed it was a big story
Sentence was placed in one of two paragraphs
- 1st was longer, more complex, and had only one editor
- 2nd was shorter, simplistic, with multiple editors
Did context affect how Ferreira & Clifton’s subjects read the sentences in their 1986 experiment (editor)? What theory does this support?
Not - no effect from context
This supports the modular approach
What was Boland trying to discover in his 1997 experiment (duck)?
Does lexical preference matter?
What did Boland do in his 1997 experiment (duck)?
Gave subjects two sentences:
- Jane saw her duck and fall down. - Jane saw her duck and some eggs.
What did Boland find in his 1997 experiment (duck)? What model did this prove?
Most people prefer the definition of “getting lower”
Most people experience Garden Paths on second sentence
Proves the Interactive Model
