| If sentence understanding is not solely determined by the surface form of a sentence and if we rarely remember the surface form of a sentence, then how can we study human memory for linguistically presented information? Verbatim recall of such materials is very bad, but we do seem to remember the gist of what was put forward. | ||
|
But what is 'gist'? We know what is it not. It is not the surface form of the sentences and it seems to include information that is not necessarily even in the surface form of the sentence. Kintsch and other researchers attempted to use ideas from linguistics, case grammar in particular, to provide a basis for representing the information that was either explicitly or implicitly present in a set of linguistic materials. The figure to the right takes one of the examples considered in your text on page 247. At the top of this figure is the input sentence, a single sentence that is quite complex syntactically. At the bottom of the figure I have provided within the dashed boxes two different expansions of this input sentence.. These expansions are intended to provide a more explicit statement of the information that is conveyed by the sentence at the top of the figure. There are two such expansions. The first is one that was created to match that provided in the text. The second is the expansion that seems most correct to me. In the center of the figure is a pictorial representation of the information (both explicit and implicit) that must be recovered from the input sentence in order to fully understand it. Notice that there are three verbs involved; form, grows, and contributes. Also note that I have connected some of the lines by dotted lines to indicate relations among the arguments. And some of the arguments are rendered in light gray to indicate that they are not explicitly mentioned. By mapping linguistic materials in this type of form, the researchers believed that they could obtain a more accurate measure of what a human subject remembered when asked to recall the input materials. This again illustrates one of the challenges that must be faced in studying the human mind. |
 |
|
| Measures that pertain only to the surface aspects of the experimental materials are not sufficient. There is a strong sense in which the empirical study of topics such as recall of linguistic materials requires a theory about mental representations in order to meaningfully measure recall performance. | ||
Understanding, Interpreting and Remembering Events |
||
| © Charles F. Schmidt | ||