Words before pictures: the role of language in biasing visual attention

Background: The present study investigated whether semantic processing of word and object primes can bias visual attention using top-down influences, even within an exogenous cueing framework. We hypothesized that real words and familiar objects would more effectively bias attentional engagement and target detection than pseudowords or pseudo-objects, as they can trigger prior knowledge to influence attention orienting and target detection. Methods: To examine this, we conducted two web-based eye-tracking experiments that ensured participants maintained central fixation on the screen during remote data collection. In Experiment 1, participants viewed a central prime—either a real word or pseudo-word—followed by a spatial cue directing them to a target on the left or right, which they located by pressing a key. Experiment 2 presented participants with real objects or pseudo-objects as primes, with primes and targets that either matched or did not match in identity. Importantly, primes in both experiments conveyed no information about target location. Results: Results from Experiment 1 indicated that real word primes were associated with faster target detection than pseudo-words. In Experiment 2, participants detected targets more quickly when primed with real objects and when prime-target identity matched. Comparisons across both experiments suggest an automatic influence of semantic knowledge on target detection and spatial attention. Discussion: These findings indicate that words can contribute to attentional capture, potentially through top-down processes, even within an exogenous cueing paradigm in which semantic processing is task-irrelevant.