<p>The power-space associations are widely discussed in the field of concept representations, with two coding systems proposed to explain this phenomenon: the verbal-spatial and visuospatial codes. However, it remains unclear whether these two coding systems are always activated in all situations, or whether they are context-dependent. To address these problems, the current study adopted two different explicit power judgment tasks, one visual and one verbal. In the visual task, all verbal cues were ruled out, and in the verbal task, all visual cues were ruled out. Each task was coupled with a secondary task—either visuospatial or verbal-spatial—to investigate the underlying activation mechanism. The results showed that both coding systems were sufficient to generate such associations separately. Most importantly, the activation of two different codes was context-dependent and interfered by the related secondary tasks.</p>