Perceptual Uncertainty and the Acquisition of Orthographic Representations

There is a high degree of flexibility in letter-position coding during visual word recognition and reading (e.g., “JUGDE” can be easily confounded with “JUDGE”). One leading account explaining this phenomenon is based on the presence of perceptual noise in the information used for locating the positions of objects—namely, letters— across space (overlap model: Gómez, Ratcliff, & Perea, 2008; Bayesian-reader model: Norris, Kinoshita, & van Casteren, 2010). This assumption is shared with formal models of visual attention (e.g., Ashby, Prinzmetal, Ivry, & Maddox, 1996; Logan, 1996). Recently, Duñabeitia, Orihuela, and Carreiras (2014) purportedly found evidence against this perceptual-noise account. They compared the performance of illiterate versus literate adults presented with sequences of four letters (or symbols). Participants were presented with a four-character string for 300 ms; this string was followed by another string that was either the same (e.g., “NDTF”-“NDTF”; “?€&<”-“?€&<”) or different (via character transposition or replacement; “NDTF”-“NTDF” vs. “NSBF”-“NTDF”; “?€&<”“?&€<” vs. “?€&<”-“?%$<”). While literate participants showed higher error rates in the transposition condition than in the replacement condition (i.e., a transposed-character effect; letters: 39% vs. 20%; symbols: 44% vs. 30%), this difference was absent in illiterate participants (letters: 62% vs. 62%; symbols: 61% vs. 57%). Duñabeitia et al. concluded that “letter-position coding is a mechanism that emerges during literacy acquisition” (p. 1275) rather than by “a generic noisy perceptual mechanism that processes all visual stimuli— letters and objects—alike” (p. 1279).