This study examined the effect of emphasis and word position on word identification by postlingually deafened adult cochlear implant (CI) listeners (n = 20). These participants performed an identification task where Swedish (quasi-) minimal pairs were drawn from sentences and presented in a carrier sentence framework. It was found that emphasised stimuli were not identified more accurately than unemphasised stimuli. A regression analysis revealed a significant main effect for words drawn from the initial position in a sentence, however there was no interaction between original word position and emphasis. Post hoc analysis of the stimuli revealed that variations in the mean intensity of items arising from their original position in the sentence or emphasis status were unlikely to account for these results. These findings have implications for those who communicate regularly with CI listeners.