This study examined the impact that metaphorical pictures and semantic transparency had on meaning recall of English idioms for L2 learners. Twenty-seven idioms of differing semantic transparency (low, mid, and high) were selected and two types of pictures were drawn for each idiom. One picture type was a literal only representation while the other was a literal+figurative representation. These idioms and picture conditions (no picture, literal only, and literal+figurative) were counterbalanced and presented to student participants (n=64) via a PowerPoint input treatment. The delayed posttests measured meaning recall of the idioms' L1 paraphrased meaning. The data showed that in absolute terms, literal-figurative pictures promoted better retention of meaning at all levels of transparency, but this finding was most robust for high-transparency idioms. However, a number of pictures, especially those with metonymical elements, led to overspecification, which has wider implications for pictorial input in general.
