Abstract
Current models attribute visual crowding to feature pooling, limited attentional resolution, and flanker substitution that compromise the identity of the target flanked by additional objects. An untested hypothesis is that the flankers may sometimes only disturb the target's perceived position, not its identity, which also causes report errors. To test this hypothesis, observers were asked to report either the central target of a three-letter string (partial report), or all letters (full report), in alternating blocks of trials. In full report, the rate of reporting the central target to the central position was comparable to that in partial report. Importantly, there was also a significant rate of reporting the correct target identity but at a wrong flanker position, supporting our hypothesis. Error analyses indicated that target-flanker position swapping and misalignment underlay observed target misplacement. In addition, target misplacement was only significant in radial, but not tangential, stimulus orientations, and could be observed at both short and long durations. Our results thus suggest a new and previously unknown crowding mechanism that interferes with the perceived position of a correctly identified target. The new mechanism, along with previously revealed ones that impair target identity, provides a more comprehensive account of visual crowding.