Winograd 1977

Reference:

Under construction

Review

Excerpts

'While the field is not yet at a stage where it is possible to lay out a precise unifying theory, this paper attempts to provide a beginning framework for studying discourse.... Its four sections attempt to:
1. Delimit the range of problems covered by the term "discourse".
2. Characterize the basic structure of natural language based on a notion of communication.
3. Propose a general approach to formalisms for describing the phenomena and building theories about them.
4. Lay out an outline of the different schemas involved in generating and comprehending language' (63).

Domains: Under construction |

Key Terms: | communication | discourse | framework | language | notion | range | structure | term | theory |



'This paper approaches the problem of studying discourse as one of understanding the cognitive structures and processes of language users. There are alternative approaches, such as text-based studies ... This paper ... proposes instead to focus on the cognitive processes of language production and comprehension. From this point of view, the text is a concrete trace of the processes, and its structure needs to be understood in terms of the processing structure ... There are clear advantages to having a framework which emphasizes the psychological processes, rather than the traces they leave, since any psycholinguistic model must deal first and foremost with the cognitive processing' (64).

Domains: Under construction |

Key Terms: | comprehension | concrete | discourse | focus | framework | language | problem | structure | text | understanding |



After considering the polar distinctions of inter-sentential versus intra-sentential, context-dependent versus context-independent, dialog versus text, and spoken versus written and dismissing them as useful because of overlap of features and considerations, Winograd asserts that 'having eliminated these simple distinctions as candidates for defining "discourse", we are left without a clear criterion for our area of study. As the rest of this paper will illustrate, there is a good deal of diversity, and a clear definition will be possible only after a larger body of theory and techniques has been developed' (66).

Domains: Under construction |

Key Terms: | definition | discourse | study | text | theory |



It is obvious that Winograd assumes the basic conduit model of communication when he says that there is a 'need to transmit meaning through a sequential medium'. Apparently the arrangement is not a particularly happy one for 'the message is forced into a linearized channel in order to be conveyed by speaking' (66).

Domains: Under construction |

Key Terms: | arrangement | communication | meaning | medium |



'The first step in building models of any cognitive process is the choice of a formalism' (72).

Domains: Under construction |

Key Terms: |



'A schema is a description of a complex object, situation, process, or structure. It is a collection of knowledge related to the concept, not a definition in the formal sense' (72).

Domains: Under construction |

Key Terms: | concept | definition | description | knowledge | schema | situation | structure |



'We can look at schemas as providing a guide for structuring the processes of production and comprehension. In the process of production, a schema ... lists the different parts and properties of a structure which must be decided upon in order to produce it.... In comprehension, the set of stored schemas is actively used in a process of "pattern recognition"' (74).

Domains: Under construction |

Key Terms: | comprehension | pattern | recognition | schema | structure |



'Psychological experiments ... have demonstrated the ways in which the application of a larger-scale or pragmatically based schema can have strong effects on the way people remember texts' (75).

Domains: Under construction |

Key Terms: | schema |



Winograd speaks of 'the schemas that form part of the cognitive structure of speaker and hearer' saying that 'the schemas can be grouped into three major areas:
1. the objects, events, and abstractions being discussed
2. the communication situation
3. the standard patterns of discourse in the language' (75).

Domains: Under construction |

Key Terms: | communication | discourse | form | hearer | language | situation | speaker | standard | structure |



'Each speaker of a language possesses a large and rather diverse set of schemas dealing with the process of natural language communication' (76).

Domains: Under construction |

Key Terms: | communication | language | speaker |



'The concepts presented in the sections above have been developed by researchers working at three different sorts of tasks: data exploration (primarily by linguists); model building (primarily in artificial intelligence); and model verification (primarily by psychologists).... Current research tends to lie in clusters along these separate lines. There is little work which combines the linguist's sophistication in recognizing the complexity of the data with the computer system builder's concern with the properties of the system as a whole, and the psychologist's demand that the resulting analysis be verifiable through experiments. If we are ever to really understand natural discourse, we have to develop methodologies which span these approaches, providing both scope and rigor in their theories' (84-6).

Domains: Under construction |

Key Terms: | analysis | data | discourse | intelligence | research | rigor | scope | system | verification |



Last Modified: July-12-96 15:46:6

Reply to randy_radney@sil.org

[A Lexicon of the Humanities | SIL Home Page | Contributions]