The present study was designed to identify and examine some of the variables that influence the focused search of semantic cases in question answering. Singer, Parbery, and Jakobson (1988) have previously reported that people can focus on the case interrogated by a question and can largely disregard irrelevant cases. In the present study, people learned facts, such as the pilot painted the garage with the roller, the spraygun, and the brush. One day later, they answered questions that focused on a particular case. For example, the question did the pilot paint with a spraygun? focuses on the instrument case. Experiment 1 revealed that people can focus on a particular case in response both to complete questions and to comparable word probes, such as "pilot spraygun." Therefore, the given-new structure of questions is not essential to focusedsearch. Experiment 2 revealed that people have a difficult time ignoring the agent case, even when it is irrelevant to the question. This corroborates proposals that agent and action information are closely interrelated in the representation of a fact. These results help to delineate the phenomenon of the focused search of semantic cases.Semantic cases playa central function in the representation of the meaning of discourse. By capturing the roles that nouns play in relation to the verbs of a sentence, such cases form an important component of the propositional idea units of a message. For example, underlying the sentence the pilot painted the fence with the brush is the proposition, (PAINT, AGENT: PILOT, PATIENT: FENCE, INSTRUMENT:BRUSH) (Kintsch, 1974). Theorists have identified numerous semantic case roles, including the agent, patient, experiencer, instrument, and location (e.g., Chafe, 1970;Fillmore, 1968;Jackendoff, 1972).There is evidence that semantic cases constitute basic perceptual-eognitive categories available to both adults and children. Adults can make comparisons among semantic cases in a manner similar to the way they compare ordinary taxonomic categories (Chaffin & Herrmann, 1984; Shafto, 1973). The language of young children reflects their growing awareness of different semantic cases (Bowerman, 1973). In addition, children can make explicit classifications based on their knowledge of such cases (Braine & Hardy, 1982;Braine & Wells, 1978).In a recent study, we examined people's ability to execute focused memory searches of semantic cases in complex facts (Singer, Parbery, & Jakobson, 1988). People learned facts such as the one expressed by Sentence 1a.Thisresearch was supported by Grant A9800 from the Natural Sciences and Engineering Research Council of Canada, and Leave Fellowship 451-86-0792 from the Social Sciences and Humanities Research Council of Canada. These studies were presented at the meeting of the Psychonomic Society, New Orleans, November, 1986. Lorna Jakobson is now at the University of Western Ontario. Requests for reprints should be addressed to Murray Singer, Department of Psychology, University of Manitoba, Winnipeg R3T 2N2, Canada.Later, ...