Adult age invariance in sentence unitization

We examined age differences on a letter detection task that was performed on four-word sentences in order to examine how letter-level and word-level processing is integrated with sentence-level unitization. Sentence-level unitization is defined as the formation of a sentence-level code that delays access to letter-level or word-level codes. There was a consistent word-frequency disadvantage for both age groups in which it took longer to detect letters within high and medium high frequency words than in low frequency words. This finding suggested that word-level and letter-level input channels were involved in a Parallel "horse race" to the central processor while simultaneously outputting codes to the sentence-level at the second tier of processing. The present data also revealed that both age groups showed a larger relative increase in reaction time for letter detection for the fourth word position compared to the third word position on syntactically "intact" sentences relative to syntactically "scrambled" sentences. These data indicated that both age groups formed sentence-level codes that made letter-level codes more difficult to access. Finally, older adults' data showed a larger cost than younger adults' data for scrambled sentences than for intact sentences. These results suggest that older adults are more reliant on a syntactical processor that facilitates the parafoveal preview during the reading of syntactically intact sentences. These data suggest that whereas there are age differences in the perceptual processing of letters and words, sentence unitization and syntactical processing remain intact with increasing adult age.