Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Copy of Copy of brainways

No description

Soojin KIM

on 15 October 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Copy of Copy of brainways

Getting Ahead by Prediction
Rolf A. Zwaan and David N.Rapp
-summarized and presented by
discourse genre
linguistic cues
background knowledge
situation models
Discourse Genres
 Readers can rely on their knowledge about narrative subgenres
to build strategies that may facilitate comprehension.
Lexical and syntactic cues often work in conjunction with each other and with background knowledge.
"Paul and Markus are going swimming today.
I hope he has his water wings."
who's 'he'?
1) first-mentioned concept
2) background knowledge
Linguistic Cues
Lexical cues
Connectives: how particular association to be built between linguistic units - “therefore, “and then,” “but,” and “however”

Pronoun referent: anaphora provide information on how incoming information should be integrated with the active memory representation

Synonyms, Direct / indirect markers
First mentions - convey the main idea
Attract reader attention
convey information about the topic and why it is important
Provide introductory material
 Receives privileged status during comprehension, and sets the stage for the encoding and retrieval of subsequent information.

Given-new material cues
Given information refers to what is already known.
new information is comprehended in light of associations and relations with the old, given information.
 Cues for indicating given-new material range from explicit reminders to changes in intonation.
Linguistic cues directly influence the ways that readers process and comprehend discourses (i.e., activation of concepts)
- explicit: “now I am going to describe what you should know for the test”
- implicit: “I saw a student enter the lab,”
“I saw the student enter the lab,”
“I saw this student enter the lab.” (Gernsbacher & Shroyer, 1989)

 Linguistic cues often function to help comprehenders integrate incoming information.
Structural cues
Particular text types can lead readers to process texts differently.

Reader-initiated process
Expectations about genre influence how readers process and remember texts (Zwaan, 1991, 1993, 1994)

These findings are not simply a function of deeper processing; readers approached these texts in qualitatively different ways that influenced the types of representations that they formed during reading. (e.g., Wolfe, 2005)

 Genre information can be both explicitly (bookstores or library organization) and implicitly (title, potential readership, availability, price, and use of terminology) provided.
Discourse Genres
Particular text types can lead readers to process texts differently.

Reader-initiated process
Expectations about genre influence how readers process and remember texts (Zwaan, 1991, 1993, 1994)

These findings are not simply a function of deeper processing; readers approached these texts in qualitatively different ways that influenced the types of representations that they formed during reading. (e.g., Wolfe, 2005)

 Genre information can be both explicitly (bookstores or library organization) and implicitly (title, potential readership, availability, price, and use of terminology) provided.
why matter?
please read
the passage
No topic group
Topic Before
Topic After
Bransford & Johnson (1972,1973)
1. background knowledge facilitates

2. to do so, it must be activated.
2 points in "Washing Clotsh"
how affect?
How background knowledge is organized in long-term memory?
semantic network (Collins & Quillian, 1969)
A concept can be represented as a node in a network and the links between them.
Frame (Minsky, 1975) or Script (Schank & Abelson, 1977)
- more structured knowledge defined by higher-order organization
Donald and his third wife entered Mario’s. When they were seated, Donald declared himself unhappy with his table, which, located near the door, was not conducive to a romantic conversation. He requested a table in the back of the restaurant. He studied the menu and ordered ossobucco for two. The food was great. Instead of a dessert, they ordered cappuccino. Donald left a big tip.

Who seated Donald and his wife,
Who took his order?
Who prepared the food?
Who delivered it to his table?
And who received the tip?
scripts might be overly rigid, given that stories are rarely if ever told if they completely follow the script. We usually feel compelled to tell a story if the events somehow violoate a script.
Neural Network models (Rumelhart, McClelland, & The PDP Research Group, 1986)
knowledge is acquired by the system as it processes input and receives feedback on its performance.
Kintsch's Construction-Integration (CI model) (Kintsch, 1988, 1998)
construction phase
integration phase
"During the earthquake, the mint collapsed"
how retrieved?
- Knowledge is retrieved via an inferencing process.
Information is usually thought of as being activated automatically, at least passively. (O’Brien, Rizzella, Albrecht, & Halleran, 1998)

Inference generation is viewed as knowledge activation or integration.
- types of inferences (Van den Brock, 1994)
Connective (or bridging) inferences
: provide a way of connecting two successive text statements; “Murray poured water on the bonfire. The fire went out.”
Elaborative inference:
is the activation of knowledge that augments the mental representation of the described situation (but not needed to integrate)
Instrument inferences
: “John let the tomato soup cool off for a while. Then he ate it.” (John used the spoon.)
Predicative inferences:
“The tired speaker finished his talk and walked over to a chair.” (the speaker is going to sit.)
Strong evidence that
comprehenders generate connective inferences
(e.g., Graesser, Singer, & Trabasso. 1994)

Evidence for more
elaborative inferences
is mixed depending on several factors;

Contextual constraints,
Individual differences,
Processing strategies,
Task instruction,
Reader expectation,
Reader preferences … influence the construction and application of inferences

The extent to which comprehension involves the
active construction
mental representation or a
more passive form of knowledge activation

Comprehension is a very active process, akin to reasoning, in which comprehender makes a conscious effort to generate bridging and elaborative inferences.
Background knowledge is retrieved automatically as a function of the processing of incoming stimuli; the comprehender passively activates background knowledge.

Neither view is completely supported by the data;
Some state that comprehenders do not indiscriminately generate inferences, but only those that are relevant, whereas others focus on the activation of information that is “easily available” through passive memory processes.
Another issue:
One issue:
event-indexing model
And the last...
Two important general intuitions about discourse comprehension;

Comprehension often seems incomplete, regardless of whether it involves automatic activation or strategic processing. (i.e. often failure of inconsistencies in texts  “sloppy” process. )

Language comprehension (narratives) often produces a sense of experiential richness.  Comprehenders can clearly become engaged in their discourse experiences. (i.e. Harry Potter books)
The role of background knowledge in the comprehension process cannot be overstated. Comprehenders rely on it, even in cases for which it is incorrect or may create problems. (students’ misconceptions are resistant to change, even in the face of refuting evidence)
 Thus, background knowledge from LTM serves as the scaffold for newly encountered information regardless of the validity of that knowledge.
Motivation (intentionality)
(Zwaan & Radvansky, 1998)
Readers attempt to integrate each incoming event with the current mental representation in working memory on each of the five dimensions.
The event-indexing model serves as a guide for determining factors that influence both the structure and content of memory for discourse.
These assumptions are entirely consistent with more general theories of language processing.
How spatial features of descriptions influence memory for texts.
Across multiple studies, evidence supports the spatial distance effect; participants take longer to identify objects that are further away from the currently described situation or events in the text.

--> Information in the same spatial frame (e.g. a particular room from the map) is strongly activated in memory, and shifts from that spatial frame lead to decrements in response accuracy and speed.
Evidence suggests that elaborate spatial representations are not formed unless
(a) the spatial features of a text are particularly salient or
(b) readers are specifically instructed to focus on the spatial descriptions.
Temporal components of texts have been considerably interesting because of its
ubiquitous quality.
iconicity assumption
; people assume that events are described in chronological order; "Veni, Vidi, Vici," - Julius Caesar
A discourse producer has to map nonlinear events onto a linear structure, whereas the comprehender has to recover the nonlinear temporal relations among these events from that linear structure.
Linguistic cues help guide readers' mapping
of such events into their appropriate described sequences. (time adverbials and verb tense and verb aspect)
Recent work of examining the
role of verb aspect
in comprehension shows that the temporal nature of a verb or verb clause
influences the accessibility of the described events from memory
. (i.e. completed events < ongoing events)
By investigating how people resolve pronouns, researchers have gained insight into how comprehenders track characters and objects in discourse.
i.e., readers tend to resolve an ambiguous pronoun using the main character of a story; thus, important characters in narratives often receive privileged status and thus are readily available during comprehension.
Comprehenders also encode the attributes and characteristics of characters as they read.
i.e., participants take longer to read outcomes that are inconsistent with emotions, behaviors, and traits suggested by preceding story contexts.
Yet, representations of characters are relatively superficial as implicit inferences in general may not be specified at levels akin to that for explicit information.
The dimension of motivation has been assessed with respect to the goals of characters in discourse.
Character goals provide justification for character behaviors and guide story plot. Thus, the extent to which readers track goals in texts is critical to theories that assess narrative comprehension.
Comprehenders appear to actively track goals; Thus,
the goals a character has for a particular situation are actively maintained in memory over the course of a reading experience.
The situation described in texts often lead readers to expect other events, or to attribute causes to particular events as a function of background knowledge.
Linguistic cues (i.e., "because" or "so") have an impact on the unfolding comprehension process and the resulting memory representation.
The comprehender's background knowledge is instrumental to forming casual representations.
"Murray poured water on the bonfire. The bonfire went out"
--> Causality is often instantiated through conceptual associations with space, goal, time, and so on.
The immersed experiencer
Readers actually experience information in a narrative as if they were participating in the activity.
Evidence from cognitive neuroscience research;
we understand the actions of others by covertly simulating these actions using the motor programs we ourselves would use to perform the action- "mirror system".

--> Like action understanding, language understanding might involve the simulation of described actions. (Rizzolatti & Arbib, 1998)
initial evidence
(Trucker & Ellis, 2004)
Participants were presented with pictures or words denoting objects and then asked to judge weather the objects were natural or manmade.
Power grip responses were faster to pictures and words denoting objects requiring power grips.
They made their judgement by manipulating an input device that required either a power grip or a precision grip.
--> This finding suggests that words make available the affordances (Gibson, 1986) of their referent objects.
Additional evidences
(Glenberg & Kaschak, 2002)
Participants listened to sentences such as "He opened the drawer, " and were asked to make sensibility judgements.
The judgements necessitated button press activities requiring either movement toward or movement away from the responder's body.
--> An action-compatibility effect was obtained such that responses were faster when the physical response was in the same direction as the movement implied by the sentence.
Motor regions of the brain are active during the comprehension of action words and sentences.
The areas of activation in the premotor cortex were somatotopically organized, such that sentences about mouth actions, hand actions, and leg actions each activated different areas.
(Hauk et al., 2004; Tettamanti et al., 2005)
These findings imply a close coupling between language comprehension and motor programs, which is consistent with the notion of comprehension as
and also consistent with
"embodied cognition"
"embodied cognition"
Cognition is grounded in perception and action and relies on the use of perceptual and motor representations, rather than of abstract, amodal, and arbitrary mental representations such as propositional networks or feature lists. (e.g., Barsalou, 1999; Glenberg, 1997)
A great deal of empirical evidences; visual representations are routinely activated during language comprehension including
object shape, orientation, and motion direction
against 'building elaborate propositional networks by activating large numbers of inferences'.

A view that assumes
perceptual and motor representations underlie cognition
provides a way to address the dilemma.
"He put the wallpaper on the table. Then he put his mug of coffee on the wallpaper."
"He put the wallpaper on the wall. Then he put his mug of coffee on the wallpaper."
Comprehension engages perceptual and motor systems by activating previous experiences, or
experiential traces
(Zwaan, 1999a, 2004)
Discourse comprehension is an essential and complex human endeavor involving processes and mechanisms associated with general cognition (e.g., memory and attention)
In a race between an elephant and a giraffe, who do you think might win?
How do you get an elephant into a refrigerator?
How do you get a giraffe into a refrigerator?
All of the animals are going to a meeting held by the king of the jungle. Only one animal does not come. Which one is it?
How do you get across a river where dozens of crocodiles live?
Categorization by:

Discourse topic
Delivery system
Author goals and intentions
`The riddles indicate that comprehension involves the construction and application of an integrated mental representation of the described events.
Linguistic cues also provide critical information that can either facilitate or hinder comprehension. (i.e., 'a' or 'the')
the goal of the chapter is to examine how processes of memory and language influence general comprehension.
genre-specific knowledge
background knowledge
What is more important than these separable components is the interplay between them, which yields a mental representation of the described situation, called "mental model" or "situation model"
Full transcript