It has been widely assumed that emotional avoidance during bereavement leads to either prolonged grief, delayed grief, or delayed somatic symptoms. To test this view, as well as a contrasting adaptive hypothesis, emotional avoidance was measured 6 months after a conjugal loss as negative verbal-autonomic response dissociation (low self-rated negative emotion coupled with heightened cardiovascular activity) and compared with grief measured at 6 and 14 months. The negative dissociation score evidenced reliability and validity but did not evidence the assumed link to severe grief. Rather, consistent with the adaptive hypothesis, negative dissociation at 6 months was associated with minimal grief symptoms across 14 months. Negative dissociation scores were also linked to initially high levels of somatic symptoms, which dropped to a low level by 14 months. Possible explanations for the initial cost and long-term adaptive quality of emotional avoidance during bereavement, as well as implications and limitations of the findings, are discussed.