Empirical research since the year 2000 on trauma and autobiographical memory in adults is reviewed and related to four enduring controversies in the field: Whether traumatic memories are inherently different from other types of autobiographical memory; whether memory for trauma is better or worse than memory for non-traumatic events; whether traumas can be forgotten and then recalled later in life; and whether special mechanisms such as repression or dissociation are required to account for any such forgetting. The review concludes that trauma and non-trauma memories differ substantially, but only in clinical and not in healthy populations. Whereas involuntary memory is enhanced in clinical populations, voluntary memory is likely to be fragmented, disorganised, and incomplete. Progress in experimental and neuroimaging research will depend on analysing how task performance is affected by the interaction of voluntary and involuntary memory and by individual tendencies to respond to trauma with increased arousal versus dissociation.