Language directs spatial attention differently in explicit and implicit tasks.

How do words with either explicit or implicit spatial meanings (e.g., DOWN, BOOT) shift our attention? Recent studies, presenting prime words followed by probe targets, suggested that, for implicit spatial words, both the spatial meaning of prime words and the target locations must be processed to induce congruency benefits. Here we examined the functional necessity of the latter location component. 91 healthy adults discriminated target letters that followed explicit or implicit spatial words. Words either did or did not have to be semantically processed. Target discrimination speed was used to compute congruency benefits. With explicit prime words, spatial congruency effects emerged without semantic processing instructions. In contrast, with implicit prime words, only instructing their semantic processing ensured a congruency benefit. This shows that, for implicit spatial words, spatial processing of target locations is not necessary; instead, processing the spatial connotation of the prime, together with the identity of the target, can induce congruency benefits. Our results help to understand previous conflicting findings.