Correlational studies suggest that clarity in teaching plays a crucial role in student learning and satisfaction from instruction. Other quantitative studies identify low-inference teacher behaviors that are components of clear instruction. The present study used qualitative methods to examine unclarity in teaching of a physics undergraduate course for non-science majors. Teaching clarity was measured both as a high- and low-inference behavior. Evidence of several data sources converge and reveal very low level of understanding of the material presented, strong dissatisfaction with instruction, and a good match between high- and low-inference teacher clarity behaviors. This is explained by the instructor's insufficient pedagogical knowledge, by his inability to apply those pedagogical principles he does know in actual classroom instruction, by his misconceptions regarding teaching and student learning, and by his detachment from his students and the subsequent lack of adjustment of instruction to them.