Words carry considerable information, but much of that information is not relevant in context. Research has shown that readers selectively activate and remember relevant information associated with words in different contexts, but it is not known when in processing this selection occurs. This experiment investigated whether context can change which properties are initially retrieved, using a speed-accuracy trade-off paradigm. Readers had to verify a property of a modifier-noun phrase (e.g., in the sentence Boiled celery is soft) within a specified interval, from 300-3,000 msec after presentation. Results revealed that properties associated with the noun alone were activated sooner than were properties that required integration of the modifier with the noun. Thus, context did not serve to influence the initial retrieval of properties but only to activate or suppress properties later in processing.
Content words communicate not only what might be thought of as "linguistic" or definitional information, but also conceptual information about the entities referred to. If you expressed surprise when someone told you, "Chris licked me," the speaker might explain, "Chris is my dog." This response perfectly explains Chris's surprising behavior, even though no one could claim that the word dog includes in its definition or purely linguistic representation the information that dogs lick people. Essentially, anything known about dogs can be referred to by dog. Given that words provide access to vast amounts of conceptual information (Murphy, 2002, chap. 11), how do comprehenders know what information is relevant in a given context? Clearly, one cannot retrieve all known facts about a word every time the word is encountered. Even if one could, most of the facts would be irrelevant to the present purposes and so would impair rather than aid comprehension. Not surprisingly, researchers have found that people tend to selectively access conceptual information, depending on the context.
Tabossi and Johnson-Laird (1980) presented words within context sentences such as (1) and (2) that emphasized one of two different aspects of the word's meaning.
(1) The goldsmith cut the glass with the diamond.
(2) The mirror dispersed the light from the diamond.
After reading one of these sentences, the participants then judged a sentence that tested an aspect of the meaning of diamond evoked by (1) or (2), for example, Diamonds are hard or Diamonds are brilliant. Verification was faster when the test sentence queried an aspect of meaning emphasized in the context sentence (see also McKoon & Ratcliff, 1988; Tabossi, 1982).
Research on text memory has also emphasized contextual influences on word interpretation. Barclay, Bransford, Franks, McCarrell, and Nitsch (1974) presented readers with sentences such as (3) or (4).
(3) The man lifted the piano.
(4) The man tuned the piano.
On a later memory test, they showed that "something heavy" was a better memory cue for (3) than was "something with a nice sound." However, the reverse was true for (4), suggesting that one sentence caused readers to retrieve the weight of pianos and the other their sound.
Context-sensitive retrieval of information is important to comprehension: One need not retrieve everything one knows about diamonds to understand sentence ( 1 ), but the sentence would be difficult to understand if one failed to retrieve the fact that diamonds are hard. Although the context sensitivity of lexical information is well agreed upon, it is not known how retrieval activates the relevant information associated with a word and not contextually irrelevant information.
One possibility is that the other words in the sentence independently prime the relevant aspects of meaning, and so the priming from the context and the target word sum to result in an active property. However, independent priming cannot explain the …