It now seems extraordinary to recall the time when clinicians believed that young people could not become depressed. Yet, it was only in the 1980s that the contemporary concept of depression entered mainstream child and adolescent psychiatry. Three decades later and this special issue of the Journal asks a variant of this question about BPD. It does say something about the crisis of legitimacy for BPD in mainstream psychiatry that we are still asking if BPD is non-normative in young people, when a wealth of evidence now suggests that BPD can take its rightful place as a severe mental disorder across the life course. 1 This series of papers systematically asks whether young people with BPD differ from their healthy peers on a range of parameters. It should not be surprising to learn that the answer is an emphatic yes. These data add to what is now a coherent literature demonstrating that young people with BPD differ from their healthy peers, not only on the parameters presented in this special issue but also on many others, such as substance use. 2 Zanarini and colleagues 3 note that psychiatrically healthy adolescents are not devoid of BPD features. They conceive of these features as 'manifestations of adolescent angst'. However, the normative adolescent development literature has tended to move away from such ideas. It might be that these are 'subthreshold' features, which point to the dimensional nature of the BPD construct. This does raise the issue of whether the arbitrary categorical threshold set by DSM-5 (five of nine criteria) is the appropriate cut point for treatment initiation, which brings us to the paper by Stepp and colleagues. 4 This begins to address this question in new analyses from the informative high-risk community sample in the Pittsburgh Girls Study. This study identifies dimensions of childhood temperament and psychopathology symptom severity that predict 'conversion' to a positive screen for BPD over a 14-year follow-up. This study highlights the need for a broad-based, integrated model of disorder 'onset', rather than one for each diagnostic 'silo'.A key premise of Stepp and Lazarus' study seems to be that childhood temperament and psychopathology are qualitatively different to BPD. Yet, existing research highlights substantial overlap in the content covered across the domains of temperament, normal-range child personality and child personality pathology. 5 This suggests that the same phenomena (in this case, emotion dysregulation and impulsivity) might be being measured but given different labels during different developmental periods. Nonetheless, this