The debate around the construct validity of complex posttraumatic stress disorder (CPTSD) has begun to examine whether CPTSD diverges from posttraumatic stress disorder (PTSD) when it co‐occurs with the diagnosis of borderline personality disorder (BPD). The present study (a) examined the construct validity of CPTSD through a latent class analysis of a non–treatment‐seeking sample of young trauma‐exposed adults and (b) characterized each class in terms of trauma characteristics, social emotions (e.g., shame, guilt, blame), and interpersonal functioning. A total of 23 dichotomized survey items were chosen to represent the symptoms of PTSD, CPTSD, and BPD and administered to 197 trauma‐exposed participants. Fit statistics compared models with 2–4 latent classes. The four‐class model showed the best fit statistics and clinical interpretability. Classes included a “high PTSD+CPTSD+BPD” class, characterized by high‐level endorsement of all symptoms for the three diagnoses; a “moderate PTSD+CPTSD+BPD” class, characterized by endorsement of some symptoms across all three diagnoses; a “PTSD” class, characterized by endorsement of the ICD‐11 PTSD criteria; and a “healthy” class, characterized by low symptom endorsement overall. Pairwise comparisons showed individuals in the high PTSD+CPTSD+BPD class to have the highest levels of psychological distress, traumatic event history, adverse childhood experiences, and PTSD symptoms. Shame was the only social emotion to significantly differ between the classes, p = .002, η² = .16. The findings diverge from the literature, indicating an overlap of PTSD, CPTSD, and BPD symptoms in a non–treatment‐seeking community sample. Further, shame may be a central emotion that differentiates between presentation severities following trauma exposure.