This data attempt to analyze different aspects of language comprehension. From a psychological point of view, the reader of a report, in extracting information from it, would be aiming to create in his or her own mind a mental representation of those events. That representation would not be as rich as those of the firsthand witnesses. Language is the mediun by which the information about the events would be transmited, indeed, without language this information could not be transmitted in such detalled form. We do claim that it is not the reader 's purpose to construct such representation; instead that purpose is to construct a mental representation of what the text is about: a sequence ef events, a place, a psycholinguistic theory, and so on. These representation are called mental models. .
Bransford's three ideas about text comprehension.
The first of these ideas was that the mental representation of a text does not correspond to any of its linguistic representation. We use syntactic and compositional semantic rules to work out the meaning of the phrases and clauses of a text. We do not typically retain detalled syntactic or semantic representation of what we hear or read. In most cases , detalled information about the form of a text desappears rapidly although we can commit it to memory (with difficulty). .
Bransford 's second idea was that comprehension is an integrative process. Syntactically based theories of comprehension inevitably focus on individual sentences. Indeed, such theories often seemed to suggest that a text representation is simple a concatenation of sentence representations. Referential links between expresions, such as the table and it , have to be established , as do links of other kinds, such as causal relations between the events, states, and processes described. A coherent text does not consist of ramdomly ordered packages of information. For the most parts, continuity of content can be assumed and used to guide other integrative processes.