LEXICAL-LEVEL AND SENTENCE-LEVEL PROCESSES IN COMPREHENSION OF WRITTEN SENTENCES

Two experiments and three measures of lexical skills (word finding, ra pid automatized naming, receptive vocabulary) were administered to 24 third and 24 fifth graders. Each experiment included two between-subje ct variables (grade and sex) and four within-subject variables (mode o f sentence presentation - one word and one-sentence-at-a-time; word ty pe - orthographic, phonological and semantic foils; position of foil i n sentence - beginning, middle, and end; and sentence type - meaningfu l and nonsense). The experiments contrasted on reading comprehension t ask: receptive sentence acceptability (Experiment 1) or expressive sen tence reproduction (Experiment 2). One-word-at-a-time presentation, wh ich highlights the lexical unit, resulted in better accuracy on the ex pressive task that required verbatim reproduction than on the forced c hoice receptive task. Word finding and rapid automatized naming that r equire active production of phonologically-referenced name codes, but not receptive vocabulary, were correlated with measures of word recogn ition and reading comprehension in the experiments and on the psychome tric tests. These results are consistent with Perfetti's contention th at a phonologically-based name code is the heart of lexical access.