Single-letter (but not multi-letter) flankers produce a leftward asymmetry in the flanker lexical decision task

We conducted three flanker lexical decision experiments to test whether multi-letter and single-letter flankers to the left and right of a target word produce symmetrical facilitation relative to an unrelated control. We also simulated these experiments using the PONG model (Snell, J. (2025). PONG: A computational model of visual word recognition through bihemispheric activation. Psychological Review, 132(3), 505–527), a model that provides quantitative predictions for this task. In Experiment 1 we found symmetrical facilitation from left-consistent and right-consistent bigrams. In contrast, Experiments 2 and 3, using single-letter flankers, revealed leftward benefits, even when the consonant–vowel distribution of the flanker letters was balanced (Experiment 3). The PONG model quantitatively captured the pattern in Experiment 1 but failed to capture the leftward advantage in Experiments 2–3. This dissociation reveals that parafoveal information during visual word recognition is hemifield-symmetric for multi-letter chunks but left-weighted for single letters, suggesting that models of visual word recognition require initial-letter or hemifield weighting to account for single-letter flanker effects.