Stimulus-response learning is traditionally attributed to repeatedly perceiving a stimulus and expressing a response. Subsequently, the perception of the stimulus can trigger the response automatically (i.e., efficiently, without an in situ conscious intention). With the present studies, we investigated whether processing verbal information that contained a stimulus and a response can result in comparable stimulus-response learning (including the automaticity consequences). In three online studies (N ~ 1000), participants were presented with "action-effect" instructions that included the information that pressing a specific button (e.g., left; action) will turn the screen background blue (stimulus/effect). Subsequently, participants categorized letters into vowels and consonants by pressing a left or right button. In some trials, the "effect" (blue background) appeared together with the target letter. Consistent with our hypothesis, the categorization responses show a response-compatibility effect: In trials with the effect (blue background) present, there are fewer errors when the required response and action-effect-instruction response overlapped than when they did not overlap. This effect is not present in the control trials (grey background), and there is no evidence for a speed-accuracy trade-off with response times. Additional preliminary data indicates some modulation of this verbally induced compatibility effect by the stimulus-response order, accompanied color information, and the degree of the relationship between the stimulus and response in the verbal instructions. In sum, albeit never directly experienced in that combination, verbal instructions that include a stimulus and a response can lead to associative learning with the consequence that perceiving the stimulus automatically prepares the respective response.