Different encoding activities during list learning, such as writing down or reading aloud, have a differential effect on memory performance. It has been argued that intermodal processing, that is, auditory processing of visually presented material and vice versa, results in a better performance than intramodal processing. This has been referred to as the "translation hypothesis." In this study, we set out to test the translation hypothesis looking at all four possible experimental conditions using visual and auditory presentation and writing and vocalization as encoding activities. The results show a similar memory performance in all conditions apart from the one in which visually presented words had to be written down. That is, in the only condition in which subjects did not hear the words (either via auditory presentation or via their own vocalization), fewer words were remembered. These findings do not support the translation hypothesis and are more in agreement with previous theoretical proposals regarding long-term modality effects.