Lexical and sublexical processes in the perception of transposed-letter anagrams

Abstract

Evidence from priming and lexical decision tasks suggests that nonwords created by transposing adjacent letter pairs (TL nonwords) are very effective in activating lexical representations of their base words, because the process of orthographic matching tolerates minor changes in letter position. However, this account disregards the possible role of sublexical processing in reading. TL nonwords are perceptually ambiguous, with lexical and sublexical processing giving rise to conflicting interpretations. The consequences of this ambiguity were investigated in a lexical decision experiment with primes that were either high or low bigram frequency TL versions of target words. Priming effects were much larger for low BF primes (e.g., pucnh–PUNCH) than for high BF primes (e.g., panit–PAINT). This finding is interpreted as evidence that lexical activation can be inhibited by competing output resulting from sublexical processing of TL letter string. We conclude that phonological processing is an important determinant of responses to TL stimuli, and we consider how this interpretation might be accommodated within the dual-route cascaded (DRC) model of word recognition.

Keywords:

• Reading
• Orthography
• Phonology
• Lexical decision
• Visual word recognition

The authors would like to thank Steve Lupker and Manolo Perea for their helpful comments on an earlier draft of this manuscript.

Notes

¹Transformation of RT data to z scores reduces the impact of missing data on error variances in the ANOVA. This is particularly important in the analysis by items in a counterbalanced design such as the one used here, when each cell in the analysis is the mean RT for a particular target/prime combination, and a subset of subjects each provides a single data point. The variation in mean RT across subjects is large in relation to the effects being analysed (for the 40 participants in this experiment, mean RT ranged from 464–695 ms for word targets). Exclusion of data for trials with incorrect responses means that individual cell means tend to be systematically increased or reduced when a “fast” or “slow” subject makes an error on a particular item. In other words, item variance is inflated by subject variance. The extent of this problem can be reduced by using standardized RT scores in the analysis of variance. For smaller scale analysis, this is less of an issue. In the present experiment, the critical comparison between TL primes with high and low bigram frequency was significant when the untransformed RT data were analysed by subjects, t(39) = 2.69, p = .01, and by items, t(114) = 2.03, p < .05.