Lawrie et al's editorial on the ‘continuum of psychosis’ is timely and welcome. ^{Reference Lawrie, Hall, McIntosh, Owens and Johnstone1} I see this debate two ways: as a doctor needing order to help ease suffering, I agree that it is better, for the time being, to keep existing diagnostic categories of psychiatric disorder, however imperfect they may be. As a patient, I of course want care, but I also want to be understood. Many psychiatrists now consider that too much of life is branded ‘disorder’: in this, none of us diminishes the reality of suffering, but we do look for better ways of explaining it. Certain scientists may hate this – but people's lives do have narrative. I think we underestimate humankind if we say that we cannot accept symptom-based descriptions of suffering. I hope I am not wrong to suggest that most of the treatments used today to improve mental health are not disease specific, but rather act on either mood, thought or both.

Nevertheless, I agree that the cry for a spectrum approach to psychosis is premature and it does not fit with my experience of so many troubled lives encountered. Peter Tyrer is correct to raise the potential problems, both clinical and pragmatic, of premature abandonment of current diagnostic classifications. ^{Reference Tyrer2} However, there remains a need to reconsider the neo-Kraepelinian model, if only to bring greater alignment with the technology that Lawrie et al hope will be to our greater mental good. It is my belief that, under the present classification system, neurobiological research cannot fully address complexity. My own view is that we have given too much attention to what Steven Rose ^{Reference Rose3} has termed ‘neurogenetic determinism’ rather than applying biological research to life (we should not risk losing the baby with the bath water, however dirty).

I would contest the presentation of the neurobiology literature as presented by Lawrie et al in the opening paragraph of their editorial. I would also contest the claim, attributed to a paper by Tandon et al, ^{Reference Tandon, Keshavan and Nasrallah4} that ‘advances in our understanding of aetiology and pathogenesis [of psychosis are] based on highly replicable neurobiological differences’. I have read that paper several times, but found, for all the studies and indeed all the words, neither one simple biomarker of any utility nor indeed anything even approaching specificity. Perhaps we should ask why this may be? Could it be that categories, clinically practicable, and needed for now, do not match the complex epigenesis of psychosis?

In concluding, I would suggest that we do not forget history. James Clerk Maxwell was bold enough to stop looking for matter and to consider the energy fields that now govern our lives and, indeed, technology that has been to our collective good. Do we need another Maxwell moment, scientifically brilliant, religion free, willing to see matters as simple as possible, but not simpler? I have no such moment to offer. But brilliant folk like Lawrie and his colleagues have that tradition and they perhaps raise the chances that such scientific inspiration can help us once again.