Journal of Experimental Psychology: Human Learning and Memory 1977, Vol. 3, No. 5, 525-538 Abstraction of Linguistic, Imaginal, and Pictorial Ideas Denise Cortis Park and William B. Whitten II State University of New York at Albany Bransford and Franks have reported that if subjects are presented with a series of related acquisition sentences that contain one, two, and three ideas, they are most likely to recognize four-idea sentences as having occurred be- fore. Bransford and Franks accounted for this finding by suggesting that subjects stored a schematic representation of experience in memory, rather than a copy of experience itself. In the current series of studies, subjects were presented with sentences that they were either to rehearse or construct into images, or they were presented with pictures of real-world scenes to remem- ber, in a Bransford and Franks paradigm. The most interesting finding was that subjects in picture conditions did not exhibit high recognition ratings for four-idea items relative to ONES, TWOS, and THREES, but subjects in sentence conditions (rehearsal and image) did. These findings were the same for Experiments 1 and 2, which differed in stimulus presentation time of acquisition items. The authors related these findings to possible differences in encoding of pictures compared to sentences and also considered the possi- bility that functional differences between picture and sentence representations exist in memory. The notion that what is stored in memory consists of an integrated version of com- ponents of experience, rather than experi- ence itself, was suggested by Bartlett in 1932. He noted that while a subject's suc- cessive recalls of a folk tale changed con- siderably over time, there were also certain elements of the tale that the subject included Bransford and Franks (1971), in a semi- nal study, supported Bartlett's conclusion that the representation of ideas in memory was abstract rather than exact. They created four complex sentences. One such sentence was "The ants in the kitchen ate the sweet jelly which was on the table." Each complex sentence was broken down into four simple in every recall attempt. Bartlett suggested ideas or ONES. The ONES were concatenated that subjects stored a schema of the major elements of the folk tale in their memories and constructed a recall protocol from this abstracted version of experience. The authors thank Robert E. Hicks of the De- partment of Psychology at the State University of New York at Albany for his statistical sug- gestions regarding this article. The authors also wish to thank Deborah Allen, Michael Frank, William B. Park, and Pamela Raskin for serving as models in the pictorial idea sets. Requests for reprints should be sent to Denise Cortis Park, who is now at the University of North Carolina at Charlotte, Department of Psy- chology, Charlotte, North Carolina 28223. An ex- amination copy of the pictorial stimuli used in this study is also available from the first author upon request. to form four TWOS and three THREES. The original complex sentence was labeled a FOUR, and the entire set of sentences de- rived from the complex sentence was called an idea set. ONES, TWOS, and THREES from each of the four idea sets were read aloud to subjects during an acquisition phase of the experiment. After each sentence, sub- jects named bars of color on a card for S sec and were then asked an elliptical question about the sentence just read. After this acquisition phase, subjects were given an unexpected recognition test consisting of presented and nonpresented ONES, TWOS, THREES, and FOURS, as well as several NON- CASES. NON CASES were sentences that com- bined ideas across idea sets or changed the meaning within an idea set. Each subject 525