Probleem 2 - Language Comprehension

Question 1

Local structure

Answer 1

Relationship between individual sentences in the discourse.

Question 2

Global structure

Answer 2

Our knowledge of the structure corresponding to the sentences together, that enables us to comprehend and remember.

Question 3

Discourse

Answer 3

Units of language longer than a sentence.

Question 4

Coherence

Answer 4

The degree to which different parts of a text are connected to each other.

Question 5

Cohesion

Answer 5

The range of possibilities that exist for linking something with what has gone before.

There are different types of cohesion:

• Reference: A semantic relation whereby information needed for interpretation in found elsewhere in text (pronominal = she, demonstrative = that, comparative = same).
• Substitution: Replacing one lexical item for another as an alternative for repeating the first.
• Ellipsis: Substituting a phrase with nothing (ex. I wish I had more talent. My sister has lot more than I do.).
• Conjunction: Expressing a relationship by using conjunctions (and/or).

Question 6

Lexical cohesion

Answer 6

A tie is made between one sentence and another by virtue of the lexical relationships between certain words in the sentence.

• Reiteration: Same word, other article&raquo_space; A boy, the boy.
• Synonymy: Other word&raquo_space; A boy, the lad.
• Hyponymy: Other broader word&raquo_space; A boy, A child.

Question 7

Anaphoric reference

Answer 7

When current expression in related to one encountered earlier.

• Anaphor = Referring expression.
• Antecedent = Previous referent.

To understand a simple pair of sentences, we must hold the antecedent in working memory long enough to link it with the anaphor. Long distance between the anaphor and antecedent impose a burden on working memory and ultimately disrupt comprehension.

Question 8

Cataphoric reference

Answer 8

Referring to a point forward.

Ex. This is how you do it. You let the herbs dry and then grind them.

Question 9

Given information vs new information

Answer 9

Given information: Information the author assumes the reader already knows.

New information: Information the comprehender is assumed to not know.

Question 10

Given/new strategy

Answer 10

Strategy used to establish coherence.

The process of understanding a sentence in discourse consists of 3 stages:

1. Identifying the given and new information in the current sentence.
2. Finding an antecedent.
3. Attaching the new info to this spot in memory.

Question 11

Direct matching

Answer 11

Strategy used to establish coherence.

Matching of underlying concepts previously introduced to the discourse.

Question 12

Bridging

Answer 12

Strategy used to establish coherence.

When there is not a direct antecedent, but can still match the sentences.

Ex. Last Christmas Eugene went to a lot of parties. This Christmas he got very drunk again.&raquo_space; Implies drunkness before, so a bridging inference is made.

Target sentences that require bridging take longer to comprehend.

Question 13

Reinstating old information

Answer 13

Strategy used to establish coherence.

When a sentence refers to someone or something which is not longer in foreground, the comprehender must reinstate the information that has te be matched with the new info.

Reinstatements increase comprehension time.

Question 14

What is the role of the working memory in language comprehension?

Answer 14

Individual differences in working memory may influence how we comprehend discourse.

The limited resources of working memory are allocated te processing certain tasks as well aas to temporarily storing the results of these tasks. When a task has considerable storage and processing demands, we may be unable to perform both functions satisfactorily.

Question 15

Reading span task

Answer 15

Examines working memory and their trade-off. Participants read aloud a series of sentences (processing) and then recall the final word in each sentence (memory).

Reading span = number of final words recalled.

Individuals with smaller reading spans had smaller working memory capacity, which made it difficult for them to comprehend references more than a few sentences back.

Question 16

Memory for discourse

Answer 16

Exists on three levels:

1. Surface representation: Remembering the exact words encountered.
   - Best remembered when it is distinctive and easily separable from the rest of the course.
2. Propositional representation: Specifies the meaning apart from the exact words.
   - Better retained than surface representations.
   - Include the meaning of presented information along with any interferences drawn before.
3. Situation model: A model of the state of affairs in the world.
   - As we comprehend the propositions of a text, we construct a mental or situational model of the world as described by the text.

Question 17

Schema/schemata

Answer 17

A structure in semantic memory that specifies the general or expected arrangement of a body of information.

We must activate the appropriate schemata in the right time to properly comprehend a story.

Information central to the schema is very well remembered while other items seem to be misplaced.

There is much evidence of influence of schemata on discourse comprehension.

Question 18

Sentence memory: A constructive versus interpretive approach.

Answer 18

Goal: Demonstrate some of the possible pitfalls of the interpretive approach to memory.

Constructive approach = Sentences are seen as info which can construct semantic descriptions of situations. These descriptions contain more info than the given sentences.

Interpretive approach = Sentence memory is a function of the deep structural information underlying the input sentences.

Experiment 1
Designed to contrast the interpretive and constructive theory by comparing recognition memory for sentences which have identical deep structures but differ in the semantic description suggested.

Two general types of sentences were used:

1. Potential Inference (PI) sentences, which suggests a potential inference based on one’s knowledge of spatial relations.
2. Non-inference (NI), did not suggest an inference.

Results:

• PI and NI sentences are remembered differently.
• In PI form inability to differentiate between old and new sentence.
• Is support for constructive theory.

Experiment 2
Investigated memory recall by PI and NI sentences. Pronoun memory was relatively good for NI sentences, but PI sentences produced very poor memory for pronoun words.&raquo_space; Is support for constructive approach.

Overall conclusion

• Recognition was shown to be rather a semantic description than an input string.
• All three experiments supported the constructive approach.

Constructive approach = top-down processing.

Question 19

Surface structure vs deep structure of a sentence

Answer 19

Surface structure: Characterizes its phonological shape.

Deep structure: Characterizes its meaning.

Question 20

Language comprehenders mentally represent the shapes of objects

Answer 20

Study examined the prediction that people activate perceptual symbols during language comprehension.

Experiment 1
Each participant saw a sentence + picture, this was either a match or a mismatch. Followed by the question has this object been mentioned?

When picture matched implied orientation&raquo_space; faster response.

Experiment 2
The participants had to name the picture after reading the sentence. Also a neutral condition was added. To explore for a facilitation or inhibition effect.

If facilitation&raquo_space; neutral/mismatch equal.
If inhibition&raquo_space; neutral/match equal.

Results:

• Result time mismatch longer.
• Neutral condition felt in between match and mismatch condition.

General conclusion:
People do activate perceptual symbols during language comprehension&raquo_space; even when orientation is implied in stead of stated.

Question 21

Perceptual symbols theory

Answer 21

Assumes that people activate and manipulate perceptual symbols during language comprehension, such that an object’s implied orientation in a sentence would be part of the mental representation of that sentence.

Question 22

Amodal theory

Answer 22

Assumes that propositional representation of an object’s orientation would not be represented.

Question 23

When peanuts fall in love: N400 evidence for the power of discourse

Answer 23

Goal:
Tracking interplay between global discourse-level factors and local anomaly constraints by means of ERPs.

Experiment 1
Participants listened to a conversation between either a person and a person or a person and a non animate object. Subjects encountered anomaly violations in the first, third and fifth sentence. Does a cartoon like setting eliminate the interpretive problems due to animacy violation?

Results: Inanimate nouns larger N400. The inanimate effect disappeared in the cartoon like setting (smaller N400). = Support for one-step model.

Experiment 2
Goal was to examine whether context appropriateness can outweigh the joint effects of anomaly and real-world plausibility.
Subjects listened to cartoon like stories with in the fifth sentence either a context-inappropriate predicate (salted) or context-appropriate predicate (in love).

Results: Context-inappropriate predicate had larger N400. So contextual appropriateness can not only neutralize animacy violance, but can even divide more preferred sentences.

Question 24

Semantics

Answer 24

Context-free combination of words within a sentence.

Question 25

Pragmatics

Answer 25

Contributions made by wider context.

Question 26

N400

Answer 26

ERP inversely related to ease of relation to semantic context. Ex. A clock is not animate&raquo_space; larger peak.

Question 27

Mental models contribute to foregrounding during text comprehension.

Answer 27

Goal:

• Demonstrate that mental models are operative during comprehension.
• To find most salient feature that characterizes mental models as mental representations and to determine whether these models are active.
• Examining the role of mental models in foregrounding concepts during text comprehension.

Experiment 1
Subjects read paragraphs. Each paragraph starts with a setting sentence which introduces main character. Second sentence was a critical sentence and included a target noun.

In the associated condition: the critical sentence described an event in which character and object were spatially associated.

In the disassociated condition: character and object were spatially disassociated.

After reading critical sentence, subject has to say whether noun is recognized as occurring in the passage. Response time was measured.

Results:

• Faster response times in associated condition.
• Response time interacted with delay; after delay of a sentence significant difference between associated and dissociated.
• Results confirm mental model approach to comprehension.

Experiment 2
Tested for speed up in reading via lexical decision task (word vs non-word). No differences found.

Experiment 3
Subjects had to refer where ‘it’ refers to. Results: associated condition faster response time.

General conclusion:
Spatial structure does influence comprehension processes&raquo_space; so it is a mental model.

Question 28

Model

Answer 28

What text is about, instead of direct representation.

Question 29

Foregrounding

Answer 29

Keeping certain tokens activated to check for conferential when a referring word is found.

Question 30

The activation of modality specific representations during discourse processing.

Answer 30

Main goal of current study was to investigate whether the activation of modality-specific representation occurs during discourse comprehension in reading conditions with no task other than to comprehend.

Participants read extended narratives while their brain activity was recorded with fMRI.

Experiment 1
Clauses were identified in the text that elicited auditory imagery, visual imagery and motor imagery and used to predict brain activity during reading.

Reading text units that imply motor action should be associated with premotor cortex etc.

Results
Auditory imagery&raquo_space; Activation temporal
Motor imagery&raquo_space; Activation motor cortex
No significant effect found for visual imagery.

Results are consistent with proposal that readers activate modality-specific representations of event features when comprehending discourse.

Experiment 2
Does motorcortex activates while reading disconnected sentences?
Story vs scrambled condition.

Results: None of the regions from study 1 showed significant effect in scrambled condition.

This suggests that readers activate modality-specific representations more strongly during coherent discourse comprehension.

Question 31

Sensorimotor simulation theories

Answer 31

Argue that the representational format of situation models is large part sensory and motor.