Transcript lecture 18 - Illinois State University Department of Psychology

PSY 369: Psycholinguistics
Language Comprehension:
Compositional meaning
Overview of comprehension
Semantic &
pragmatic
Syntactic analysis
analysis
Word
recognition
We’ve talked about the syntax here
Language
How does the meaning combine?
perception
Input
He sent a letter about it to Galileo, the great Italian scientist.
Overview of comprehension
Semantic &
pragmatic
Syntactic analysis
analysis
Word
recognition
How does the meaning combine?
Language
How do we represent sentence meaning?
perception
Input
He sent a letter about it to Galileo, the great Italian scientist.
Meaning beyond the word

Move to compositional semantics

Not all meaning resides at the level of the
individual words.



Conceptual combinations
Meaning of Sentences
Meaning across multiple sentences (texts & discourse)
Conceptual combination

How do we combine words and concepts?

We can use known concepts to create new ones

Noun-Noun combinations


Modifier noun
Head noun

“Squirrel box”

“Skunk squirrel”

“Helicopter flower”
Conceptual combination

How do we combine words and concepts?

Relational combination



Property mapping combination



Relation given between head and modifier
“squirrel box” a box that contains a squirrel
Property of modifier attributed to head
“skunk squirrel” a squirrel with a white stripe on its back
Hybrid combinations


A cross between the head and modifier
“helicopter flower” a bird that has parts of helicopters and
parts of flowers
Conceptual combination

How do we combine words and concepts?

How do the theories do?


Instance theory has problems
Separate Prototypes? (big wooden spoon)



But sometimes the combination has a prototypical feature that is
not typical of either noun individually (pet birds live in cages, but
neither pets nor birds do)
Modification? (brown apple, peeled apple)
Extending relational structure between characteristics?


When nouns are “alignable” (zebra horse)
But non-alignable nouns are combined using a different
mechanism (zebra house)
Meaning beyond the word

Move to compositional semantics

Not all meaning resides at the level of the
individual words.



Conceptual combinations
Meaning of Sentences
Meaning across multiple sentences (texts & discourse)
Memory for sentences
Think back to the beginning of class. Which of the following
sentences did you read?

Galileo, the great Italian scientist, sent him a letter about it.

He sent Galileo, the great Italian scientist, a letter about it.

He sent a letter about it to Galileo, the great Italian scientist.

A letter about it was sent to Galileo, the great Italian scientist.
Memory for sentences
Sachs (1967, 1974)
 Heard (read):


Tested:





“He sent a letter about it to Galileo, the great Italian scientist.”
Same: He sent a letter about it to Galileo, the great Italian scientist.
Act/Pass: A letter about it was sent to Galileo, the great Italian scientist.
Formal: He sent Galileo, the great Italian scientist, a letter about it.
Meaning: Galileo, the great Italian scientist, sent him a letter about it.
Measured accuracy of detecting changes
Memory for sentences
Percent correct
Sachs (1967, 1974)
Sachs (1967)
semantic
change
active/passive
change
formal change
100
90
80
70
60
50
identical
sentence
0
80
160
Amount of interpolated material
(number of syllables)
Conclusions: Sentence meaning gets preserved, surface
structure (and syntax) forgotten
Propositions

How do we represent sentence meaning?

Propositions



Two or more concepts (arguments) with a relationship
(relations) between them
 Arguments – particular times, places, people,
objects, etc. (nouns)
 Relations - May be used for any kind (e.g., actions,
attributes, positions, class memberships)
Smallest unit of knowledge that can be judged as true or
false
Complex sentences consist of combinations of smaller
propositional units
Propositions

How do we represent sentence meaning?

Propositions

Two or more concepts with a relationship between them
A mouse bit a cat
bit (mouse, cat)
mouse
agent

patient
relation
bit
Can represent this within a network framework
cat
Deriving Propositions

More complex example:

Children who are slow eat bread that is cold



Slow children
Children eat bread
Bread is cold
Past
Slow
Children
Eat
Bread
Cold
Evidence for propositions


Bransford and Franks (1971, 1972)
Ratcliff and McKoon (1978)
Evidence for propositions


Bransford and Franks (1971, 1972)
Study-Recognition Test Task

Read sets of sentences, answered a question about each, later
presented sentences and asked whether they were new (not
previously presented) or old (previously presented)
The girl broke the window on the porch. Broke what?
The hill was steep.
What was?
The cat, running from the barking dog, jumped on the table. From what?
The tree was tall.
Was what?
The old car climbed the hill.
Did what?
The cat running from the dog jumped on the table.
Where?
The girl who lives next door broke the window on the porch. Lives where?
…
Evidence for propositions


Bransford and Franks (1971, 1972)
Study-Recognition Test Task
All of the sentence came from 4 complex sentences. The full
complex sentences were not presented at study.
e.g., The girl who lives next door broke the large window on the porch
…
The girl lives next door.
The girl broke the window.
The window was on the porch.
The window was large.
Evidence for propositions


Bransford and Franks (1971, 1972)
Study-Recognition Test Task
Test:
Old - same sentences that were presented at study
New - based on the propositions in the complex sentence, but not
presented at study (including the full complex sentences)
Noncase - based on new propositions not based on the complex
sentences (mixing of propositions across the different situations)
Evidence for propositions

Recognition confidence

Yes 5
Yes 4
Yes 3
Yes 2
Yes 1
0
No 1
No 2
No 3
No 4
No 5
Bransford and Franks (1971, 1972)
Study-Recognition Test Task






Results:
•
new
old


False recognition of sentences
that they were not previously
presented with
•
Accurate rejections of noncases
(different propositions)
•
Unable to distinguish between the
old and new cases that came
from the same complex
sentences


★
fours threes twos
ones
# of propositions
noncases
Evidence for propositions


Bransford and Franks (1971, 1972)
Study-Recognition Test Task
Recognition confidence
Conclusions:
Yes 5
Yes 4
Yes 3
Yes 2
Yes 1
0
No 1
No 2
No 3
No 4
No 5






•
new
old
•




★
fours threes twos
ones
# of propositions
noncases
Participants remembered the basic
meaning (propositions)
Participants spontaneously combined
the propositions into larger units
Evidence for propositions

Ratcliff and McKoon (1978)
Tested 3 hypotheses:
1.
2.
3.
Sentences stored as single unit
Sentences stored as connected propositions
Sentences stored verbatim
Evidence for propositions


Ratcliff and McKoon (1978)
Study-Recognition Test Task


Read sets of 4 unrelated sentences, then presented words (one at a
time) and asked whether the words were in the preceding sentences
Dependent Measure: Priming - manipulated the order of the words at
test
The mausoleum that enshrined the tsar overlooked the square.
The clutch failed to engage.
The beggar forgave injustice but resented hunger.
Satire hurt the incumbent.
hunger Y Saturn N square Y mausoleum Y beetle N
Evidence for propositions

Ratcliff and McKoon (1978)

Involves two propositions:


P1 [OVERLOOK, MAUSOLEUM, SQUARE]
P2 [ENSHRINE, MAUSOLEUM, TSAR].
The mausoleum that enshrined the tsar overlooked the square.
The clutch failed to engage.
The beggar forgave injustice but resented hunger.
Satire hurt the incumbent.
Evidence for propositions


Ratcliff and McKoon (1978)
Predictions (if Hypothesis 2: propositions are the memory representation):


If prime word from the same sentence, then should respond faster
If prime word from the same proposition, then should respond faster
than if from a different proposition (within the same sentence)
The mausoleum that enshrined the tsar overlooked the square.
The clutch failed to engage.
square
clutch
square
tsar
square
mausoleum
Across sentences
Between two propositions
in the same sentence
Within a single
proposition
Evidence for propositions


Ratcliff and McKoon (1978)
Predictions (if Hypothesis 2: propositions are the memory representation):


If prime word from the same sentence, then should respond faster
If prime word from the same proposition, then should respond faster
than if from a different proposition (within the same sentence)
**91 msec

Results
671
square
clutch
Across sentences
**111 msec
580
square
tsar
560
square
mausoleum
Between two propositions
in the same sentence
Within a single
proposition
Evidence for propositions


Ratcliff and McKoon (1978)
Predictions (if Hypothesis 2: propositions are the memory representation):


If prime word from the same sentence, then should respond faster
If prime word from the same proposition, then should respond faster
than if from a different proposition (within the same sentence)
**20 msec

Results
671
square
clutch
Across sentences
580
square
tsar
560
square
mausoleum
Between two propositions
in the same sentence
Within a single
proposition
Evidence for propositions


Predictions (if Hypothesis 2: propositions are the memory representation):



Ratcliff and McKoon (1978)
If prime word from the same sentence, then should respond faster
If prime word from the same proposition, then should respond faster
than if from a different proposition (within the same sentence)
Conclusions

Support the hypothesis that propositions are used to organize our
memories of sentences
Inference in comprehension

Not all propositions come from the bottom-up

Elaboration - integration of new information with
information from long term memory


Memory for the new information improves as it is
integrated
Inferences - a proposition (or other representation) drawn
by the comprehender

From LTM, not directly from the input
Inference in comprehension

Bransford, and colleagues (1972, 73)


We draw inferences in the course of
understanding new events.
The inferences get encoded into our memory
of the events.

e.g., drawing inferences of instruments
Inference in comprehension

Bransford, and colleagues (1972, 73)
Saw (or heard):
John was trying to fix the birdhouse. He was looking for the nail when his father
came out to watch him and to help him do the work.
Tested:
John was using the hammer to fix the birdhouse when his father came out to
watch him and to help him do the work.
Result:
was not mentioned in the text,
but was inferred
Participants falsely believed that they had heard this sentence
So memory is not only of propositions in the original sentence, but may also
include additional propositions that may have been inferred
Arguments against propositions

Propositions are symbolic and amodal

Referential problem:


Implementation problem:


Has been very difficult to develop a propositional parser
Lack of scientific productivity:


Disconnected with outside world (symbols referring to
other symbols)
More work on what you can do with propositions than is
there evidence of the psychological reality of propositions
Lack of a biological foundation:

How do biological (or neurological) data constrain
propositions
More than propositions

Barclay (1973)




Subjects are presented with sequences of sentences that create a
spatial array, like:

The bear is to the left of the moose.

The moose is to the right of the lion.

The moose is to the left of the cow.

The lion is to the left of the bear.
Array: lion < bear < moose < cow
Subjects are asked either to remember the sentences or to
remember the order
Afterwards, people asked to remember the array also ‘remember’
sentences they didn’t actually hear, such as:
 The bear is to the left of the cow
(also faster to verify, Potts, 1974)
More than propositions

Bransford, Barclay, and Franks (1972)
Hear: There is a tree with a box beside it, and a chair is on top
of the box. The box is to the right of the tree. The tree is green
and extremely tall.
(a): The tree is to the left of the chair.
(b): The chair is to the left of the tree.
Recognition Task Result: correctly rejected (b) but accepted (a)
Mental
model
Mental Models
Hear: There is a tree with a box beside it, and a chair is on top
of the box. The box is to the right of the tree. The tree is green
and extremely tall.
• These experiments suggested that contexts are not
simply lists of propositions, but that these propositions
are somehow ‘merged’ to create `world-like’
representations
• Johnson-Laird (1983): While processing, humans construct
representations of worlds/situations related (identical with?)
those built from perception
Discourse in memory

Kintsch’s model


The Construction-Integration Model
Discourse occurs in a series of cycles


In each cycle



As each sentence comes in it gets integrated into the
discourse
Construction phase - activate relevant concepts
Integration phase - keep only the most relevant
elaborations
Multiple levels of representation formed

Surface form, textbase (propositional), situation model
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface form
S
N
VP
V
Jack scanned
NP
the
newspaper
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface
form
Textbase
S
N
VP
Examine
V
Jack
Jack scanned
NP Newspaper
the
newspaper
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface
formSituational Model
Textbase
S
N
VP
Examine
V
Jack
Jack scanned
NP Newspaper
the
newspaper
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Similar meaning
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
If Better
memory here
Evidence for
surface form
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Adds inference
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Infers which section did he scan.
If Better
memory
here
Evidence for
Strong textbase
Discourse in memory

Kintch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

consistent
inconsistent
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Consistent with situation model.
Evidence for
Strong situation model
If Better
memory
here
Discourse in memory
Kintch and colleagues (1990)
1.2
1
Trace strength

0.8
Situational
model
Textbase
0.6
0.4
Surface form
0.2
0
-0.2
0 min
40 min 2 days 4 days
Delay
Embodiment in language

Embodied Representations

Many researchers assume that cognition is “embodied” (or
“grounded”) rather than “abstract” (e.g., Barsalou, 2008)


Activates representations associated with the body and actions
Theoretical proposals from many disciplines




Linguistics: Lakoff, Langacker, Talmy
Neuroscience: Damasio, Edelman
Cognitive psychology: Barsalou, Gibbs, Glenberg,
MacWhinney, Zwaan
Computer science: Steels, Feldman
Embodiment in language

Embodied Representations

Much of this work argues that language is embodied
(e.g., Barsalou, 2008; Glenberg, 2008; Zwaan & Taylor, 2006)
 Perceptual and motor systems play a central role in language
production and comprehension (and meaning/concepts)

Words and sentences are usually grounded to perceptual,
motoric, and emotional experiences.

In absence of immediate sensory-motor referents, words and
sentences refer to mental models or simulations of
experience
Embodiment in language
Embodied Representations



Simulation hypothesis (Gallese, 2008)
Simulation exploits some of the same neural structures activated
during performance, perception, imagining, memory…
 Language gives us enough information to simulate
 Processing (producing or comprehending) walk involves the
use of representations involved in the act of walking
producing or
comprehending
“walk”
Embodiment in language

Evidence for Embodied Representations


Stanfied & Zwaan (2001)
Presented participants with sentences
John put the pencil in the cup.
John put the pencil in the drawer


See a picture and ask “does this describe what you read
about?”
Results: faster at saying horizontal pencil with drawer and
vertical pencil with cup
Embodiment in language

Evidence for Embodied Representations

Zwaan et al (2004)
Presented participants with a sentence

A: The pitcher hurled the softball at you.
B: You hurled the softball at the pitcher.
See two pictures and ask “are these pictures the same object”

B
A

Results: faster at saying ‘Yes’ when sentence matched the
pictures (e.g., sentence A and pictures in A, if the ball is small
and then gets big, it is coming towards you)
Embodiment in language

Evidence for Embodied Representations


Hauk et al (2004)
Do action words activate the motor cortex? fMRI study
50 words from 3 semantic subcategories
(words matched for freq, length, imageability, etc.)
Rated for whether words reminded them of face, arm, or leg

Movement Comparison: moved their foot, finger, or tongue
Embodiment in language

Evidence for Embodied Representations



Hauk et al (2004)
Do action words activate the motor cortex? fMRI study
Action words did activate some of the same areas as the
movements
Summing up

Traditional

Embodiment of Meaning

Cognition = Computation


Representation by
propositions


Propositions are abstract
relations

Cognition is serving perception
and actions
Representation = Patterns of
possible bodily interactions
with the world (lawfully related
to the world)
What an object, event,
sentence means for you, is
what you can do with the
object, event, sentence.
Meaning beyond the word

Move to compositional semantics

Not all meaning resides at the level of the
individual words.



Conceptual combinations
Meaning of Sentences
Meaning across multiple sentences (texts & discourse)
Discourse Psycholinguistics

Traditional Psycholinguistics


Determining what happens when we understand
sentences
Broader View

How we resolve/understand sentences against the
current discourse representation

Sentence comprehension is a process that anchors the
interpretation of the sentence to the representation of the
prior text
Discourse Psycholinguistics

Traditional Psycholinguistics


Determining what happens when we understand sentences
Broader View

How we resolve/understand sentences against the current
discourse representation

Sentence comprehension is a process that anchors the
interpretation of the sentence to the representation of the prior
text
Processing Discourse

What is discourse?

The ways that we process (i.e., comprehend and
remember) units of language larger than a sentence


Lectures, personal narratives, expository discourse
Units of analysis larger than a sentence

Applies to both spoken and written forms
Discourse processing is sort of like syntactic
processing – a way of organizing/connecting
the different pieces in to larger chunks. Here
the chunks are larger than sentences.
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

To whom does “him” refer to?
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

To whom does “him” refer?
Bach
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

To whom does this “him” refer?
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

To whom does this “him” refer?
Bach again
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

To whom does this “him” refer?
Bach again
Why not Abe?
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

Huh!?
Bill and Ted traveled through time and space.
Bill asked, “Dude, did you find him?”
“Yeah, Bach was in the music store, but I haven’t found Abe
yet. Dude, you should hear him play, he rocks!”
“Excellent! Man, we’ve got to get these dudes back to
school before we get there.”

Huh!?
Oh yeah, they’re time travelers.
Characteristics of Discourse

Local Structure (microstructure):

The relationship between individual sentences



Cohesion
Coherence
Global Structure (macrostructure):

The relationship between the sentences and our
knowledge of the world
Characteristics of Discourse

Local Structure (microstructure):

The relationship between individual sentences



Cohesion
Coherence
Global Structure (macrostructure):

The relationship between the sentences and our
knowledge of the world
Characteristics of Discourse

Local Structure (microstructure):

The relationship between individual sentences


Cohesion
• Does the discourse “stick together”?
• Interpretation of one sentence depends on other
sentences?
Coherence
• Does the passage make sense?
• Logical consistency and semantic continuity?
Characteristics of Discourse

Cohesion: Interpretation of one sentence
depends on other sentences

Referential Cohesion


Substitution Cohesion


“We’ve got to get these dudes back to …”
And many more


“Dude, you should hear him play…”
Ellipsis, conjunction, lexical cohesion (See pg 160 of
textbook for examples)
The relationship between the referring expression
and the antecedent create referential cohesion of
discourse
Types of Referential Cohesion

Anaphoric Reference

Using an expression to refer back to something previously
mentioned in discourse
“…Bach was in the music store …”
“Dude, you should hear him play, he rocks.”

Cataphoric Reference

Using an expression to refer forward to something that is
coming up in discourse
Dude, did you find him?”
“Yeah, Bach was in the music store...”
Comprehending Anaphoric References
Daneman and Carpenter (1980)

Task: Reading a passage and answer questions about the referents of
pronouns
Sitting with Richie, Archie, Walter and the rest of my gang in the Grill yesterday, I began to
feel uneasy. Robbie had put a dime in the juke box. It was blaring one of the latest
“Rock and Roll” favorites. I was studying, in horror, the reactions of my friends to the
music. I was especially perturbed by the expression on my best friend’s face. Wayne
looked intense and was pounding the table furiously to the beat. Now, I like most of
the things other teenage boys like. I like girls with soft blonde hair, girls with dark curly
hair, in fact all girls. I like milkshakes, football games and beach parties. I like denim
jeans, fancy T-shirts and sneakers. It is not that I dislike rock music but I think it is
supposed to be fun and not taken too seriously. And here he was, “all shook up” and
serious over the crazy music.
Question: Who was “all shook up” and serious over the music?
Comprehending Anaphoric References
Daneman and Carpenter (1980)

Task: Reading a passage and answer questions about the referents of
pronouns
Sitting with Richie, Archie, Walter and the rest of my gang in the Grill yesterday, I began to
feel uneasy. Robbie had put a dime in the juke box. It was blaring one of the latest
“Rock and Roll” favorites. I was studying, in horror, the reactions of my friends to the
music. I was especially perturbed by the expression on my best friend’s face. Wayne
looked intense and was pounding the table furiously to the beat. Now, I like most of
the things other teenage boys like. I like girls with soft blonde hair, girls with dark curly
hair, in fact all girls. I like milkshakes, football games and beach parties. I like denim
jeans, fancy T-shirts and sneakers. It is not that I dislike rock music but I think it is
supposed to be fun and not taken too seriously. And here he was, “all shook up” and
serious over the crazy music.
Question: Who was “all shook up” and serious over the music?
Comprehending Anaphoric References
Daneman and Carpenter (1980)

Task: Reading a passage and answer questions about the referents of
pronouns
Sitting with Richie, Archie, Walter and the rest of my gang in the Grill yesterday, I began to
feel uneasy. Robbie had put a dime in the juke box. It was blaring one of the latest
“Rock and Roll” favorites. I was studying, in horror, the reactions of my friends to the
music. I was especially perturbed by the expression on my best friend’s face. Wayne
looked intense and was pounding the table furiously to the beat. Now, I like most of
the things other teenage boys like. I like girls with soft blonde hair, girls with dark curly
hair, in fact all girls. I like milkshakes, football games and beach parties. I like denim
jeans, fancy T-shirts and sneakers. It is not that I dislike rock music but I think it is
supposed to be fun and not taken too seriously. And here he was, “all shook up” and
serious over the crazy music.
Question: Who was “all shook up” and serious over the music?
 Reading Span Test


Smaller reading spans = smaller working memory capacity
Manipulated how many sentences intervened between the
pronoun ‘he’ and the antecedent ‘Wayne’
Comprehending Anaphoric References
Daneman and Carpenter (1980)

Conclusions: The number of intervening sentences don’t
matter for high span people, but does for low span
Results
Percentage correct

120
100
80
60
40
20
0
small (2-3) medium (4- Large (6-7)
5)
Number of sentences
between pronoun and
antecedent
High span
medium span
Low span
Characteristics of Discourse

Coherence:

Given/new distinction


•
•
Readers expect speakers to provide cues as to what
information is old (already known by the listener) and what
is new (not known)
Making Inferences
 Filling in missing pieces of information to maintain
coherence
Haviland and Clark (1974)
Singer, Halldorson, Lear, & Andrusiak (1992)
Developing coherence
Haviland and Clark (1974)

Process of understanding a sentence in discourse
context involves 3 stages:
1. Identify the given and new info in the current sentence
2. Find an antecedent in memory for the given
information
3. Attach the new information to this spot in memory
Developing coherence
Haviland and Clark (1974)

Task: Press a button when you understand the sentence.
We got some beer out of the trunk. The beer was warm.
Developing coherence
Haviland and Clark (1974)

Task: Press a button when you understand the sentence.
We got some beer out of the trunk. The beer was warm.
Definite article “The” signals that “The beer” is given information
Developing coherence
Haviland and Clark (1974)

Task: Press a button when you understand the sentence.
We got some beer out of the trunk. The beer was warm.
Definite article “The” signals that “The beer” is given information
Connect the new information “was warm” to the appropriate
discourse concept
“some beer”
This process is called Direct Matching
Developing coherence
Haviland and Clark (1974)

knowledge
Task: Press a button when you understand theWorld
sentence.
We checked the picnic supplies. The beer was warm.
Definite article “The” signals that “The beer” is given information
Connect the new information “was warm” to the appropriate
discourse concept
“??”
Developing coherence
Haviland and Clark (1974)

knowledge
Task: Press a button when you understand theWorld
sentence.
We checked the picnic supplies. The beer was warm.
Definite article “The” signals that “The beer” is given information
Connect the new information “was warm” to the appropriate
discourse concept
“picnic supplies”
Need a bridging inference to connect “the warm
beer” to “picnic supplies”
Developing coherence
Haviland and Clark (1974)



Task: Press a button when you understand the sentence.
Conclusion: If you don’t know the old information and need to
make an inference, this may slow down comprehension.
Direct Matching
Typical results
We got some beer out of the trunk. The beer was warm.
Comprehended
faster

Bridging Inference
We checked the picnic supplies. The beer was warm.
Takes more
time
World knowledge
Developing coherence
Singer, Halldorson, Lear, & Andrusiak (1992)


Task: Press a button when you understand the sentence, if
given a question, answer Yes or No.
Conclusions:


Suggests that the bridging inference was made
More time consuming to make coherence of temporal than causal
relations
Causal condition
Requires
bridging inference
Results
“Murray poured water on the fire.”
“The fire went out.”
Faster reading time
T/F “water extinguishes fire”
Temporal condition “Murray drank a glass of water.”
No required
“The fire went out.”
inference
T/F “Does water extinguish fire?”
Faster
“T”
Brief summary

Local Structure (microstructure):



Discourse is coherent if its elements are easily related.
Coherence is achieved with cohesive ties between
sentences.
Comprehension is impeded when
 There are no antecedents, forcing a bridging inference
 The antecedent was not recent, forcing a
reinstatement of the antecedent.
Characteristics of Discourse

Local Structure (microstructure):

The relationship between individual sentences



Coherence
Cohesion
Global Structure (macrostructure):

The relationship between the sentences and our
knowledge of the world
Characteristics of Discourse

Global Structure (macrostructure):

The relationship between the sentences and our
knowledge of the world
Jill bought a new sweater. Sweaters are sometimes made of
wool. Wool production gives some farmers a good livelihood.
Farming is a high-risk business. On the news last night, I
saw a group of business executives discussing recent trends
in the stock market.

Okay local structure, but each sentence isn’t
relevant to an overall topic of discourse
Characteristics of Discourse

Read story to class (from Bartlett, 1932)
Characteristics of Discourse

Global Structure (macrostructure):

Schemas (Scripts)


General knowledge structures for common social
situations
Genres

Narrative structure


Story grammars - extension of idea of grammatical
rules, specify the organization of a story
Expository structure

Different structures
Characteristics of Discourse

Global Structure (macrostructure):

Schemas (Scripts)


General knowledge structures for common social
situations
Genres

Narrative structure


Story grammars - extension of idea of grammatical
rules, specify the organization of a story
Expository structure

Different structures
Effects of world knowledge
If the balloons pooped, the sound would not be able to
carry since everything would be too far away from the
correct floor. A closed window would also prevent the
sound from carrying since most buildings tend to be
well insulated. Since the whole operation depends on
a steady flow of electricity, a break in the middle of the
wire would also cause problems. Of course the fellow
could shout, but the human voice is not loud enough to
carry that far. An additional problem is that a string
could break on the instrument. Then there could be no
accompaniment to the message. It is clear that the
best situation would involve less distance. Then there
would be fewer potential problems. With face to face
contact, the least number of things could go wrong.

Bransford & Johnson (1972)
Effects of world knowledge
Rocky slowly got up from the mat, planning his escape. He
hesitated a moment and thought. Things were not going well.
What bothered him most was being held, especially since the
charges against him had been weak. He considered his
present situation. The lock that held him was strong but he
thought he could break it. He knew, however, that his timing
would have to be perfect.
Prison escape
OR
Wrestling match

Anderson et al (1977)
Effects of world knowledge

Schemas (Scripts)

Mental structures of how the world works,
acquired through experience

A whole package of information about what we
know about the world and events



Generic story of situations
A framework with causal information
Used to facilitate comprehension of discourse, as
well as to guide recall (and reconstruction)
Effects of world knowledge

Schemas (Scripts)

Generic story of situations
Restaurant Script
Scene 1: Enter
Scene 2: Order
Scene 3: Eat
Scene 4: Pay
Go inside
Go to table
Sit down
Get menu
Read menu
Choose food
Give order
Get food
Eat food
Ask for check
Received check
Tip waiter
Pay check
Exit
Effects of world knowledge
Bartlett (1932)

Task:



Read native American folk tale
Write down everything that you can remember
from that story that I read earlier
Bartlett had them recall after a longer periods of
time (between 15 mins. Up to 10 years later)
Effects of world knowledge
Bartlett (1932)

Results:
 Participants’ memories changed to fit their
existing beliefs (reconstructive memories)
Added new details
 Changed details
 Deleted details
Conclusions: We use our Schema to facilitate
comprehension of discourse, as well as to guide
recall (and reconstruction)


Effects of world knowledge
Invernizzi & Abouzeid (1995)


Read two European tales (cry wolf & stone soup)
2 audiences


European North American children
Ponam children (New Guinea)
Effects of world knowledge
Invernizzi & Abouzeid (1995)


Retelling of boy who cried wolf
Ponam children (New Guinea)

Once upon a time Kalai and his family they lived on an island.
Kalai’s mother always carried him everywhere. One day
Kalai’s mother and father went out fishing. Kalai’s mother said,
“Kalai, you are too small to go out fishing in the sea. You
should stay home with your grandfather.” Kalai was lonely on
the beach. Kalai said, “How could I get my family home?” He
sat down and decided to get his family home. He got his red
laplap and ran down to the beach and waved his laplap to his
family and said, “Fire, fire.” His brother saw his laplap and
went home. When they arrived they saw nothing.
Effects of world knowledge
Invernizzi & Abouzeid (1995)


Retelling of boy who cried wolf
European North American children

Kalai was running up and down the beach yelling “Fire, fire.”
Everybody came home. The next day the same thing
happened. They came home. The next day came, but the
house caught on fire. He ran up and down the beach, but
nobody came. Kalai kept waving the flag. Nobody came.
Suddenly they saw the flames and the smoke and they came,
but it was too late. Everything had burnt down to the ground,
and his brother told him if he kept telling lies that nobody will
come when you call for help.
Effects of world knowledge
Invernizzi & Abouzeid (1995)


Impact of different schemata
European North American children


Ponam children (New Guinea)
Recalled factual detail about settings, events, and
outcomes, but leaving out things like consequence,
resolution, moral (generally seemed to miss the
point)
Conclusions: We use our cultural schemas to facilitate
comprehension of discourse, as well as to guide recall


Setting, precipitating events, goal reaching
aspects, story resolutions
Effects of world knowledge
When do we use the schema? During
comprehension or recall?
Smith and Swinney (1992)


Task: presented stories (like the “balloons” one)




Results:



Collected sentence by sentence reading times
Had them recall the sentences
Some people were given a title for the story, others not
Overall, reading times were faster with a title that without
Stories with titles: More words were recalled and more
“intrusions” (details consistent with the schema but not in the
story)
Conclusions:

Schemas are used in both on-line comprehension and recall
Effects of world knowledge

Summary

We use schemas to
 Facilitate the comprehension of discourse
 To guide recall (and reconstruction)
Effects of Genre

Not all kinds of discourse follow the same
structure

Different effects, purposes, etc.

Expository discourse


Narrative discourse



Convey info about a subject (e.g., textbook, lecture)
Tell a story: Introduce characters & settings, establish
a goal, etc.
APA style
Newspaper articles
Expository Structure

Reading texts, listening to lectures, etc.


Organized with different relationships (but can still draw a tree
structure)
Relationships





Collection - ideas or events related on the basis of some commonality
Causation - ideas are joined causally so that one idea is identified as
the antecedent and another as the consequence
Response - ideas are joined in a problem/solution or question/answer
relationship
Comparison - ideas are related by pointing out similarities and
differences
Description - general ideas are explained by giving attributes or other
specific details
Narrative structure
Once there was a woman. She saw a tiger’s
cave. She wanted a tiger’s whisker. She put
food in front of the cave. The tiger came out.
She pulled out a whisker.

The story has a structure, a story grammar
Narrative structure

Story grammar - can depict with a tree structure
Story
Setting Episode
Once there was a woman.
Event
She saw a tiger’s cave.
Reaction
Goal
She wanted a tiger’s whisker.
Overt Response
Action Consequence
She put food in front of the cave.
Event
The tiger came out.
Event
She pulled out a whisker.
Narrative structure
Thorndyke (1977)

Level effect

High hierarchy statements
Lower in the hierarchy.


Read more
slowly but
are better
remembered.
Comprehensibility and recall were tied to inherent plot
structure, independent of passage content
She wanted a tiger’s whisker.
The tiger came out.
Characteristics of Discourse
Trabasso & Suh (1993)

Test to see if structure effects whether inferences are
made

Task: Think aloud task

Read through the story aloud (one sentence at a time) and
talk aloud about their understanding of that sentence
Trabasso & Suh (1993)
Sequential version
Hierarchical version
Once there was a girl named Betty.
One day, Betty found that her mother’s birthday was coming soon.
Betty really wanted to give her mother a present.
Betty went to the department store.
Betty found a pretty purse.
Betty found that everything was too expensive.
Betty bought the purse.
Betty could not buy anything.
Her mother was very happy.
Betty felt sorry.
Several days later, Betty saw her friend knitting.
Betty was good at knitting.
Betty decided to knit a sweater.
How does
this sentence connect up
Betty selected a pattern from a magazine.
withBetty
thefollowed
rest ofthe
the
story? in the article.
instructions
Finally, Betty finished a beautiful sweater.
Betty pressed the sweater.
Betty folded the sweater carefully.
Betty put it in the closet for the
Betty gave the sweater to her mother.
next time she was going out.
Her mother was excited when she
Berry was very happy.
saw the present.
Trabasso & Suh (1993)
Hierarchical version
S Once there was a girl named Betty.
E One day, Betty found that her mother’s birthday
G
A
O
O
R
E
was coming soon.
Betty really wanted to give her mother a present.
Betty went to the department store.
Betty found that everything was too expensive.
Betty could not buy anything.
Betty felt sorry.
Several days later, Betty saw her friend knitting.
S = Setting
E = Event
R = Reaction
G = Goal
O = Overt Response
A = Action
S
G
A
A
O
A
A
O
R
Betty was good at knitting.
Betty decided to knit a sweater.
Betty selected a pattern from a magazine.
Betty followed the instructions in the article.
Finally, Betty finished a beautiful sweater.
Betty pressed the sweater.
Betty folded the sweater carefully.
Betty gave the sweater to her mother.
Her mother was excited when she saw the
present.
Trabasso & Suh (1993)
Hierarchical version
S Once there was a girl named Betty.
E One day, Betty found that her mother’s birthday
G
A
O
O
R
E
was coming soon.
Betty really wanted to give her mother a present.
Betty went to the department store.
Betty found that everything was too expensive.
Betty could not buy anything.
Betty felt sorry.
Several days later, Betty saw her friend knitting.
S
G
A
A
O
A
A
O
R
Betty was good at knitting.
Betty decided to knit a sweater.
Betty selected a pattern from a magazine.
Betty followed the instructions in the article.
Finally, Betty finished a beautiful sweater.
Betty pressed the sweater.
Betty folded the sweater carefully.
Betty gave the sweater to her mother.
Her mother was excited when she saw the
present.
Is a superordinate goal that motivates the subgoal of the next episode
S
E
G
A
E
O
S
O
G
R
A
A
A
O
A
O
R
Trabasso & Suh (1993)
Sequential version
S Once there was a girl named Betty.
E One day, Betty found that her mother’s birthday
G
A
O
O
R
E
Betty was good at knitting.
Betty decided to knit a sweater.
Betty selected a pattern from a magazine.
Betty followed the instructions in the article.
Finally, Betty finished a beautiful sweater.
Betty pressed the sweater.
Betty folded the sweater carefully.
Betty put it in the closet for the next time she
was going out.
R Berry was very happy.
S
G
A
A
O
A
A
O
was coming soon.
Betty really wanted to give her mother a present.
Betty went to the department store.
Betty found a pretty purse.
Betty bought the purse.
Her mother was very happy.
Several days later, Betty saw her friend knitting.
The goal is already filled, so not related to the subgoal of the next episode
S
E
G
A
O
E
O
S
R
G
A
A
O
A
A
O
Trabasso & Suh (1993)
Results

In a think aloud task
 participants mentioned the superordinate
goal in the hierarchical condition
 but not the sequential condition

Story grammar structure matters

Strongly support the hypothesis that readers do make
global causal connections during reading.
Discourse in memory

Daily Summary:



Schemas are used to structure comprehension and
memory
Discourses have internal structures that impact
comprehension and memory
Weekly summary:

Evidence supports the psychological reality of a number of
different representations
 Propositions & propositional networks
 Embodied representations
 Inferences
 Schemata and scripts
 Situation models
Discourse in memory

Kintsch’s model


The Construction-Integration Model
Discourse occurs in a series of cycles


In each cycle



As each sentence comes in it gets integrated into the
discourse
Construction phase - activate relevant concepts
Integration phase - keep only the most relevant
elaborations
Multiple levels of representation formed

Surface form, textbase (propositional), situation model
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface form
S
N
VP
V
Jack scanned
NP
the
newspaper
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface
form
Textbase
S
N
VP
Examine
V
Jack
Jack scanned
NP Newspaper
the
newspaper
Discourse in memory

Kintsch’s model
Jack scanned the newspaper.
Surface
formSituational Model
Textbase
S
N
VP
Examine
V
Jack
Jack scanned
NP Newspaper
the
newspaper
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Similar meaning
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
If Better
memory here
Evidence for
surface form
Discourse in memory

Kintsch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

Read before
Adds inference
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Infers which section did he scan.
If Better
memory
here
Evidence for
Strong textbase
Discourse in memory

Kintch and colleagues (1990)
It was Friday night and Jack and Melissa were bored, so
they decided to catch a movie. Jack scanned the
newspaper. He saw that they could just make the nine
o’clock showing of the hot new romantic comedy. Off
they went.

consistent
inconsistent
Did this sentence occur in the paragraph?
Jack scanned the newspaper.
Jack looked through the newspaper.
Jack looked through the movie ads.
Jack looked over some editorials.
Consistent with situation model.
Evidence for
Strong situation model
If Better
memory
here
Discourse in memory
Kintch and colleagues (1990)
1.2
1
Trace strength

0.8
Situational
model
Textbase
0.6
0.4
Surface form
0.2
0
-0.2
0 min
40 min 2 days 4 days
Delay
Summary

Discourse processing is both complex and
flexible


Multiple representations
Processing depends on context
The house the handyman the teacher hired painted
burned down.

Center
embedded
structures
This sentence
is syntactically
legal.





So
is it soburned
hard to comprehend?
Thewhy
house
down.
One possibility is that there are limitations as a result of our
The
house the handyman painted burned down.
cognitive systems
Memory and comprehension

Center embedded structures

The house burned down.
Memory and comprehension

Center embedded structures


The house burned down.
The house the handyman painted burned down.
Memory and comprehension
This one may be legal, but

Center embedded structures



that doesn’t mean that it is
(easily) comprehensible
The house burned down.
The house the handyman painted burned down.
The house the handyman the teacher hired painted
burned down.

(the handyman that the teacher hired painted the house that burned
down)
Memory and comprehension
The man that the woman that the child hugged kissed laughed.
S

NP
RR
NP
C
VP

NP
VP
V
V
laughed
The man
that
RR
NP
the woman
NP
C
kissed
VP
NP
that
V
The child
hugged
Typically we build right
branching structures
But here there is a big
series of branches
down the center
Memory and comprehension
The man that the woman that the child hugged kissed laughed.

Most readers having trouble figuring out who did what to
whom (called thematic role assignment).
 Easier to assign thematic roles in the two sentences that
form it:



The man that the woman kissed laughed.
The woman that the child hugged kissed the man.
Possible explanation for the trouble:
 Insufficient working memory resources to
retain the intermediate products of
computation made building the complex
syntactic structure
Discourse in memory

Daily Summary:



Schemas are used to structure comprehension and
memory
Discourses have internal structures that impact
comprehension and memory
Weekly summary:

Evidence supports the psychological reality of a number of
different representations
 Propositions & propositional networks
 Embodied representations
 Inferences
 Schemata and scripts
 Situation models
Measuring memory span
When at last his eyes opened, there was no gleam of triumph, no shade of anger.
Measuring memory span
The taxi turned up Michigan Avenue where they had a clear view of the lake.
Measuring memory span

Recall the last words
When at last his eyes opened, there was no gleam of triumph, no shade of anger.
The taxi turned up Michigan Avenue where they had a clear view of the lake.
Measuring memory span
I turned my memories over at random like pictures in a photograph album.
Measuring memory span
I will not shock my readers by describing the cold-blooded butchery that followed.
Measuring memory span
He had an odd elongated skull which sat on his shoulder like a pear on a dish.
Measuring memory span
You can check out the books that you need for this course at the local library.
Measuring memory span
The radio station was promoting the concert with free tickets and back stage passes.
Measuring memory span
The professor could be seen on weekends in the backyard garden pulling out weeds.
Measuring memory span

Recall the last words
I turned my memories over at random like pictures in a photograph album.
I will not shock my readers by describing the cold-blooded butchery that followed.
He had an odd elongated skull which sat on his shoulder like a pear on a dish.
You can check out the books that you need for this course at the local library.
The radio station was promoting the concert with free tickets and back stage passes.
The professor could be seen on weekends in the backyard garden pulling out weeds.


Ok for two sentences; Hard at 3 sentences; Very hard
for 4 or more.
Used to classify readers as high and low span
(there is a “speaking span” version too)
Memory and online comprehension
The Capacity Theory of Comprehension

(Just & Carpenter, 1992)

Proposed that individual differences in working memory
capacity should influence how readers comprehend
sentences
Memory and online comprehension
The Capacity Theory of Comprehension

(Just & Carpenter, 1992)
Proposed that individual differences in working memory
capacity should influence how readers comprehend
sentences

The animacy of the first noun may constrain the possible
interpretation of the sentence
Semantically Unconstrained:
The defendant examined by the lawyer shocked the jury.
The defendant that was examined by the lawyer shocked the jury.
Semantically Constrained (so should be faster if animacy can be used)
The evidence examined by the lawyer shocked the jury.
The evidence that was examined by the lawyer shocked the jury.
“that was” disambiguates these
sentences

Memory and online comprehension
Just & Carpenter (1992)
600
550
The defendant examined by the
lawyer shocked the jury.
500
msec
Just the ambiguous sentences
animate NP
Inanimate NP
450
400
The evidence examined by the
lawyer shocked the jury.
350
300
Low-span
High-span
High span readers could use the semantic information
to resolve the ambiguity
Memory and online comprehension
King and Just (1991)

Verbs which could provide strong pragmatic cues as to which of
the two potential actors in the sentence was the agent:
Embedded clause
The robber that the fireman rescued stole the jewelry.


Two possible agents:


the robber
the fireman

Two verbs, which is the
main verb of the
sentence?:


rescued
stole
Memory and online comprehension
King and Just (1991)

Verbs which could provide strong pragmatic cues as to which of
the two potential actors in the sentence was the agent:
Strong bias  .The robber that the fireman rescued stole the jewelry.


No bias



The robber that the fireman rescued watched the program.
The robber that the fireman detested stole the jewelry.
The robber that the fireman detested watched the program.
Can bias which Noun goes with which Verb pragmatically (or not)
Questions:
 Can speakers use this information?
 Does memory have an impact?
Memory and online comprehension
King and Just (1991)

Verbs which could provide strong pragmatic cues as to which of
the two potential actors in the sentence was the agent:
Embedded relative verb




Main verb
The robber that the fireman rescued stole the jewelry.
The robber that the fireman rescued watched the program.
The robber that the fireman detested stole the jewelry.
The robber that the fireman detested watched the program.
Memory and online comprehension
King and Just (1991)
Comprehension accuracy


Verbs which could provide strong pragmatic cues as to which of
the two potential actors in the sentence was the agent:
Method
90
H
80 L
H
L

H
H
Data
70
L
L
60
both
Results



Word-by-word moving window
procedure
relV
main
V
neither


% correct on a T/F
comprehension question
when relative clause is tested
Reading times by region
Low-capacity subjects had lower comprehension overall & slower reading in
syntactically difficult regions
High-capacity subjects did NOT improve with pragmatic info
Low-capacity subjects did improve with pragmatic info
Memory and online comprehension
Garnsey, Pearlmutter, Pirog (2003)
The professor (who was) confronted by the student was
not ready for an argument.
The professor (had) confronted the student but was
not ready for an argument.
Question:
Do readers differ specifically in how quickly they can use
disambiguating words to rule out incorrect alternatives?
Memory and online comprehension
Garnsey, Pearlmutter, Pirog (2003)
Eye fixations were analyzed separately
- By whether preview of “by” while still fixating on verb likely
If last fix was here,
trial not used
The professor confronted by the student was not ready to …
If last fix was here,
trial coded as
Preview Unlikely
If last fix was here,
trial coded as
Preview Likely
Memory and online comprehension
Readers who score high on the Reading Span test
- Make better use of a peripherally visible disambiguating word to
quickly rule out a preferred but incorrect interpretation
Memory and online comprehension




Just & Carpenter (1992) - high span readers used
semantic information early, but low span readers didn’t
King & Just (1991) - high span readers did not use
pragmatic information to resolve ambiguity, but low span
readers did
Garnsey, Pearlmutter, Pirog (2003) - span differences may
also depend on where the eye lands (which determines
what kind of preview readers get)
What information is used to resolve syntactic
ambiguities depends on individuals working memory
capacity (but see Walters and Caplan (1996) for alternative
view)
Memory and comprehension

Brief summary so far:

What do we remember about sentences?

Syntax may not be too important later, we
remember the meaning of sentences but not so
much the form (syntax) of the sentence

What is the role of syntax in comprehension?

Syntax is important for getting the right
interpretation during on-line comprehension

Memory capacity may play an important role in
determining what kinds of information we can use
to comprehend sentences
Summing up

The results of sentence comprehension are
meaning representations


Some debate over what these representations are
Whatever they are, they get integrated with each
other and with existing knowledge from LTM