Research has demonstrated that when discrete pieces of information are integrated together at encoding--imagining two items together as a single entity, for example--there is a mnemonic benefit for their relationship. A separate body of literature has indicated that the presence of emotional information can have an impact on the binding of associated neutral details, in some cases facilitating associative binding (MacKay et al. Memory and cognition 32:474-488, 2004; Mather, perspectives on psychological science 2:33-52, 2007), and in other cases impeding the processing of associated details (Easterbrook, Psychological Review 66:183-201, 1959; Kensinger, Emotion review 1:99-113, 2009). In the present experiments, we investigated how memory for neutral words is affected by the emotionality of the information with which they are presented (whether with an emotional word or a second neutral word) and the encoding context (integrated or nonintegrated strategy). Participants viewed word pairs and were instructed to visualize the items as an integrated unit or to visualize them separately from one another. The results of Experiment 1 showed a disproportionate mnemonic benefit for neutral items that were integrated with other neutral items over those integrated with emotional items. The results of Experiments 2A and 2B showed that this effect interacted with encoding time: When given 2 s to encode, participants showed no effect of integration on memory for neutral-neutral pairs, but showed a significant mnemonic benefit for integrating emotional-neutral pairs. When given 4 or 6 s, the integrative benefit increased significantly for neutral-neutral pairs but decreased for emotional-neutral pairs. These results suggest that creating an integrated mental image of two neutral items requires a more time-consuming process than integrating an emotional and a neutral item, but that extra effort may result in a downstream mnemonic benefit.