Several studies have investigated potential interactions between emotion and memory by focusing on analyses of accuracy and response bias. These studies typically show higher accuracy and laxer response bias to emotional relative to neutral words. Fewer studies, however, have examined interactions between emotion and memory by focusing on response time data. Here, we report a recognition memory experiment in which emotional words, neutral words, and semantically related neutral words were used as probes. Participants were slower to reject novel emotional words than to reject novel neutral words, whereas they exhibited equivalent response times to emotional and neutral studied words. Quantile analysis showed that such slower rejection of emotional novel words was restricted to slower responses, suggesting potential interactions between emotion and higher order processes during recognition. These findings are interpreted in light of affective theories of exposure and theories of emotional processing.