Numerous languages, including Hindi, German, and (Old) English, have so-called correlative structures, which combine a question word with a demonstrative pronoun (e.g., 'Where X is, there you will find Y'). The structure seems to draw attention to the referent of the demonstrative ('there'). We investigated this hypothesis in a wayfinding experiment. Ninety-nine German participants saw three scenarios, each consisting of written directions and a map showing three possible destinations. For each German text, we identified a crucial sentence providing information on which direction to take at a junction, and created three versions, differing only in phrasing (canonical: 'Turn right at X', implicitly salient correlative: 'There where X is, there you must turn right', and explicitly salient: 'You must pay attention here: On seeing X, you must turn right'). Each participant was presented with one of the three phrasings in each of the three scenarios. Following the main task, participants completed a sentence-recognition task. All conditions were counterbalanced across participants and scenarios. We observed a significant effect of sentence recognition on selecting the correct destination, but phrasing did not show any significant effect. A post-hoc-analysis revealed a significant effect of phrasing on sentence recognition: Memory was best for the explicitly salient sentence. We conclude that specific phrasing can indeed draw attention to a sentence, but this is not particularly the case for correlative structures and does not seem to be strong enough to influence performance in a wayfinding task.