A transposed-word effect in same-different judgments to sequences of words

Abstract

The present study examined transposed-word effects in a same-different matching task with sequences of 5 words. The word sequences were presented one after the other, each for 400 ms, the first in lowercase and the second in uppercase. The first sequence, the reference, was either a grammatically correct sentence or a scrambled ungrammatical sequence of the same words. The second sequence, the target, was either the same as the reference or differed either by transposing the second and third words or the third and fourth words in the first sequence or by replacing the same 2 words with different words in Experiment 1 or by a single word replacement in Experiment 2. The results showed that "same" responses were easier to make with grammatically correct references and that "different" responses were harder to make when the difference involved a transposition compared with a replacement. This transposed-word effect was found to be independent of reference grammaticality in Experiment 1. Experiment 2 again found a transposed-word effect for ungrammatical sequences, but here the effect was reduced compared with grammatical sequences. The effects found with ungrammatical sequences are taken to reflect the noisy bottom-up association of word identities to locations along a line of text, and this process combines with the influence of top-down grammatical constraints when "different" judgments are harder to make.