DISCOURSE COMPREHENSION AND MEMORY

DISCOURSE COMPREHENSION AND MEMORY

What is comprehension?

Discourse comprehension is influenced by what is given the discourse (e.g., coherence, title) and by what an interpreter brings to the task (e.g., working, memory, knowledge of the world)

Discourse comprehension involves perceiving and integrating skill and integrating all kinds of information-linguistic or pragmatic, explicit or implicit contained discourse so as develop a coherent and correct picture of the events describe in the discourse. For full appreciation of discourse, an interpreter has such extra task as these: separate facts from opinions; evaluate the relevance of materials to the author’s thesis or to her own interpreting goals: appreciates the beauty, aptness or novelty of expressions: grasp the point of joke, irony or sarcasm and above all, retain at the last the main point of what has been interpreted (Taylor and Taylor 1983, p. 393-94).

Discourse processing must be studied by looking at both the interpreter and the discourse. The interpreter brings to the task; knowledge of language, knowledge of the world and of the subject matter, working memory capacity, reasoning ability, processing perspectives and so on. Many excellent readers have trouble following income tax forms and legalese. In a well structured discourse, sentences should be linked and structured, important ideal should be distinguishable from unimportant ones, implied information should be inferable, conclusion should be easily drawn, figurative language should be apt, the title should be informative.  

Listening and reading may differ at early stages of processing: listening involves an evanescent auditory image and auditory word recognition, whereas reading involves a static visual representation and visual word recognition. But a sentence and discourse levels, listening comprehension and reading comprehension become similar.  Listening skills and reading skills predict each other once word recognition skills have developed. Reading comprehension ability is indistinguishable from listening comprehension ability (Palmer, Macleod, Hunt and Davidson 1985, p.59)

The correlation between listening comprehension listening comprehension and reading comprehension though high, is not perfect partly because a listener must process speech as he hears it, whereas a reader can process written material at his own pace.

How is Comprehension studied?

For conversation, evidence of understanding and short term remembering can be seen in its smooth flow: one participant speaks appropriately if he, as the listener, has understood and remembered, partly at least, another speaker’s utterance. For written text, what kind of evidence of comprehension can we obtain? (A reader of this book may move her eyes over a page and turn pages diligently without grasping a thing!) How a written text comprehended can be tested on-line while reading is taking place.

One common experimental technique is to let a reader control the rate of presenting a stimulus, be it a word, a sentence or paragraph.

Coherence and comprehension

Cohesive Devices

Local coherence between sentences and clauses can be achieved via cohesive devices that repeat, substitute, or delete linguistics items and forms already used (Halliday & Hasan 1976). Particularly effective cohesive devices are sentence connectors (e.g., therefore, furthermore in formal discourse and so, and in informal discourse), which not only explicitly connect consecutive sentences but also indicate the nature of this connection.

Anaphora and Antecedents

In determined the antecedent and anaphora, an interpreter sometimes has to use knowledge of the world as well as, or more than, the linguistics cues available in the text. For the speaker or writer, anaphora should save time and effort by indicating that he is still on the original discourse topic, which presumably has been partially processed. However, anaphora can increase processing time and effort when antecedent is difficult to indentify.

Paragraph Structure

Successive sentences should be only linked but also around a single discourse topic, as in a paragraph found in expository discourse, often written but sometimes spoken. A paragraph consist of a topic sentences, often given at the beginning of a paragraph, and several sentences that support, e.g., elaborate, illustrate, justify – the topic sentences. The paragraph may include a concluding remark.

The topic sentence can often be identified even when the sentences in a paragraph scrambled (Plafflin 1967). In reading a series of paragraphs, subjects read topic sentences more slowly than nontopic sentences, and read sentences introducing major topic shifts more slowly than those introducing minor topics shifts (Lorch, Lorch & Mogan 1987). Apparently, readers construct a list of discourse  topics as they read, adding a new topic to the list whenever it cannot be integrated with one already included on the list (hence a long reading time).

Story Structure

A good story has four elements:

1.      Character(s) or protagonist(s) –hero(s) or heroine(s)  whose actions the story traces;

2.      Plot; a short of chart or road map by which a character’s action are traced , beginning with his goal and ending with the attainment or nonattainment of the goal with twist and turns.

3.      Setting; an environment in which character’s action takes place.

4.      Moral; concise summary of the story, often as a lesson life.

A story is also segmentable into parts, according to some story grammarians. Story grammar predicts story processing, according to story grammarians: in recalling stories after hearing them, subjects of ages tend to recall canonical stories better than noncanonical stories to the original order (e.g., Mandler 1978; Mandel and Johnson 1984).

Idea Units And Levels Of Importance

Rating and Recalling Idea units

Ideas expressed in discourse vary importance. To study the relative importance of ideas, first a narrative or expository passage has to be segmented into “idea units” which then will be rated as to their importance, because sentences, though they can be the major linguistics and processing units, are not ideal as units for the present purpose,, for they vary enormously in length and structure , from one word (“Come”) to as long as ninety words (as is this awkward sentences), a paragraph, a page, a few pages and so on, so that researchers look for units that are less variable than sentences.

Important information takes longer to read, presumably with deeper processing, than unimportant information.  When the reading time of important information was limited by an experimenter, recall f the important information was still good, presumably because of the extra cognitive effort (e.g., Intense attention , concentration) allocated to it (Britton, Muth, & Glynn 1986).

Factors Affecting Importance

The importance rating can be influenced by the perspective from which people encode (perceive and store information while listening to or reading) a story. Various factors-some inherent in discourse and some adopted by an interpreter –influence ratings of the importance of idea units. The units rated important are processed slowly and attentively and are weil recalled

Knowledge and Context

Implied Information and Inference

A writer or speaker does not spell out every conceivable item of information in her speech or writing; instead, she leaves some information unsaid or implied, in accordance with the principle of minimum information; an interpreter fills in or infers the implied information using context and knowledge of the world.   What kind of information is inferred and when it is inferred at the time of encoding or off recall? For a highly predictable consequences off a fourteenth story roof , death will be inferred  of events  such as someone falling of a fourteenth –story roof, death will be inferred only if the next sentence explicitly refers to the funeral (Singer & Ferreira 1983).

Listeners to, or readers of, a story were more likely to take inferences on its important elements-the goal , plan, and actions of characters –than on less important elements such as states (e.g., “John owned a car”) Seifert Robertson, & Black 1985). The inferences on the important elements were made during as well as after comprehension.

Title and Knowledge Structure

The kind of inferences people draw can be strongly affected by the knowledge structured activated by a little.

Context Effects

Context-both linguistics (a title or surrounding text) and situational aids discourse processing by narrowing the domain of interpretation, activating an appropriate knowledge structure. As context, even a picture plus a title can be over powered by extraordinary even vivid in memory. And a context can sometimes lead misinterpretation.

Figurative Language

Metaphors, simile, idioms, proverbs, sarcasm, parables, allegories and the like, are example of figurative language that should as  a rule be interpreted nonliterally.  Of the variety of figurative language, three types – Metaphors, Idiom and Proverb –have become favorite stimuli in psycholinguistic research. An idiom is a formulaic phrase whose meaning may not be predicted from the meaning of individual words. An idiom may have both literal and figurative meaning (e.g., pull one’s leg) or may have only figurative meaning (e.g., the apple of my eye). A proverb is the experience and observation of several ages, gathered and summed up into one expression (Oxford English Dictionary).

Proverbs can be

1.      Literal only

2.      Literal  or figurative

3.      Figurative  only

Indirect Speech

As figurative language as interpreted for more than one meaning, so is indirect speech interpreted for more than one function. According to one theory, the interpretation of an indirect request involves three steps:

1.      The listeners construct the literal meaning.

2.      Check the context for its plausibility.

3.      And if is implausible, applies a rule of conversation to derive the conveyed meaning.

But as with figurative language, direct interpretation may occur, bypassing the literal meaning, especially if the nonliteral meaning is conventional and spoken in context.

Mental Representation

To understand and remember discourse to build a representation of its meaning in memory. A mental representation, also called a mental mode or situation model, is constructed not in terms of exact words and sentence forms; rather, it is in terms of the events – actor objects and their relations-describes in the discourse (e.g., Johnson-Laird 1983)  Mental representation of discourse in terms of the events described rather than of the exact words and sentence structures. A mental representation:

1.      Is constructed, guided by such knowledge structure as script and schema.

2.      Includes words instantiated and information inferred

3.      Is concise and coherent

4.      Is centered around a characters goal in a story, or around a little in exposition

5.      May contain inaccuracies

6.      Can be in a verbal or visual  form

During Comprehension

Fletcher  (1986) compared eight strategies , four local and four global, that have been proposed by other researchers. The four strategies select:

1.      Most recent

2.      Most recent topical

3.      Most recent propositions that contains the most frequent argument

4.      Leading edge (recent and important proposition)

The four global strategies select the most recent propositions that:

1.      Follow a script.

2.      Correspondent to the major categories of a story grammar.

3.      Indicate or allow one to infer) a character’s goal or plan achieving it.

4.      Are parts of the most recent discourse topic.

After Comprehension

People reconstruct or reproduce a discourse differently, depending on instructions, such as

1.      Be accurate;

2.      Accept inference;

3.      Be compatible with the meaning of a passage.

With delays of one or seven days, compared to no delays , pieces of information distant from the original ones more likely to be accepted (Brockway, Chmielewski, & Cofer 1974)