Conceptual beliefs learners hold may influence learning strategies and outcomes of learning. Little research, however, has investigated English as a Foreign Language (EFL) learners' views of multimodality presented in the most accessible learning materials, their English textbooks. A total of sixteen 11th graders participated, half from the language gifted (LG) class and half from the regular social science (RSS) class. They were interviewed to investigate conceptions and how proficiency level regulates conceptions. The phenomenographic method was applied to iteratively analyze the students' utterances. Five qualitatively different categories of conceptions emerged, including, from the lowest in rank, "Read for textual information only," "Read with engaging visual aids," "Read bi-modally for facts," "Read bi-modally to comprehend," and "Read bi-modally and critically." The results revealed that bi-modal reading has the potential to enhance comprehension as indicated by the greatest number of conceptions, "Read bi-modally to comprehend," and motivate learning, as in "Read with engaging visual aids," ranked the second. Next, investigating the tie of proficiency to conceptions suggests that while the LG group was attuned bi-modally in advanced cognitive processing, the RSS group was more divided, in that some heavily relied on text while others were comfortable using both text and images to read critically.