Encoding the levels of information in pictures and words.

摘要: From a levels-of -processing framework (e.g., Craik & Lockhart, 1972), we derive the idea that levels of information implicit in stimulus, such as its physical configuration, name, and category to which it belongs, may become available for subsequent processing at different times after stimulus onset. In particular, tasks allow use codes should be performed more rapidly than those require "deeper" information. There are two important implications here: The first is ability code does not mean has been "matched to" representation memory. second "depth" effects (i.e., less name conceptual) demonstrable within both pictorial verbal materials. approach whole contrasted with current dual-coding approaches Paivio, Note 1), which, while they meaning imaginal symbolic systems, seem forced assume certain types cenceptual easily accommodated system. This leads prediction some kinds conceptual will directly from pictures, but must instead subject via an interaction between systems. We tested these assumptions speeded-inference task (Van Rijn, 1973), properties unambiguous interpretation reaction time differences across materials same explicit depth. Briefly, found pictures yield faster inferences do words when semantic required performance (Experiments 1 2), sooner (Experiment 4) 5). Moreover, representations proper nouns) function symbolically extent have advantage discriminability over 6 7), unless features added them redundant needs perform 8). results best interpreted levels-of-processing framework, multiple or exist activated by appropriate stimuli, rather stimuli themselves embody encoded used needed.