Dichoptic reading: The role of meaning in binocular rivalry

These experiments sought to determine whether meaning influences the predominance of one eye during binocular rivalry. In Experiment 1, observers tried to read meaningful text under conditions in which different text streams were viewed by the two eyes, a situation mimicking the classic dichotic listening paradigm. Dichoptic reading proved impossible even when the text streams were printed in different fonts or when one eye received a 5-sec advantage. Under non-rivalry conditions, the observers were able to read text presented at twice the rate used for dichoptic testing, indicating that cognitive overload does not limit performance under conditions of rivalry. In Experiment 2, observers were required to detect repeated presentations of a probe target within a string of characters presented to one eye. Although this task was easily performed under monocular viewing conditions, it proved difficult when the two eyes received dissimilar character strings. This was true regardless of whether the probed eye viewed nonsense strings, real words, or meaningful text. In a condition designed to encourage semantic processing of one eye’s view, the observers were required to detect animal names as well as to detect the probe target. Performance remained inferior to that measured under monocular conditions. Even the observer’s own name proved insufficient to influence the predominance of one eye under conditions of dichoptic stimulation. When two text strings were physically superimposed and viewed monocularly, essentially no probes were detected, indicating that the failure to see some probes during rivalry reflects a limitation unique to dichoptic viewing. These results contradict theories attributing binocular rivalry to an attentional process that operates on monocular inputs that have received refined analysis. This conclusion may be limited to rival stimuli whose meaning is defined linguistically, not structurally.