In his paper published in the June 2012 issue of World Psychiatry, J. Parnas focuses on the fundamental disturbance in patients before the onset of schizophrenic illness 1. He refers to the ICD-8, in which a personality disturbance was considered to be the clinical core marker of schizophrenia. In the ICD-9, this concept was retained including a disturbance of the will (inertia and negativism). In his editorial in the same issue 2, M. Maj calls for follow-up studies on Parnas' Gestalt approach in order to capture the basis of schizophrenia. This dialog immediately brought to my mind the work of another Danish psychiatrist, E. Dein (1922–1975). In the 1950s, Dein performed follow-up studies in patients within the spectrum of schizophrenia 3. He defined autism as “a permanent and socially perceptible deviation as to the interpersonal relations of the individual, characterized by a pervasive tendency to keep an abnormal distance to other people, to adopt a passive attitude and to be emotionally indifferent” 4. He found that 109 of 142 patients with chronic schizophrenia had autism before the onset of their illness, compared to only one of 52 patients with episodic schizophreniform (probably schizoaffective) states 3. The results obtained by Dein confirmed Kraepelin's observation that autism is a clinical core marker of schizophrenia. The Gestalt approach proposed by J. Parnas is a major contribution in this respect. During our validation study on depression and mania scales 5, Dein told me that the most severe case of contact disturbance he had ever observed was in a patient with manic-depressive illness. In that illness, however, the disturbance does not appear outside acute episodes, in contrast to schizophrenia, where it represents a personality trait 4. In his editorial, M. Maj also makes some remarks on the practical problems of Parnas' model. Among these is that the core Gestalt is possibly very reliable in the hands of super experts, but dangerously volatile in ordinary clinical practice 2. In this respect, Dein's words 4 echo Maj's concerns: “It is true that the ability of psychiatrists to perceive subtle shades in contact varies considerably, and perhaps a natural wish to establish contact may lure many psychiatrists into overestimating their rapport with the patient. It is at any rate inexpedient if a criterion so decisive for the diagnosis as autism must depend too much on the individuality of the psychiatrist, and it really may become a problem if the amount of his experience and competence comes to play a dominant role in clinical discussion”. Dein also states that “when evaluating possible autism… one should… try to close one's eyes to other symptoms, as for instance disturbances of thinking, transitivistic phenomena, delusions or hallucinations; these symptoms are sometimes so suggestive of the diagnosis of schizophrenia that one may be prejudiced in favour of the presence of autism” 4. When making an attempt to close the eyes of psychiatrists of the DSM-IV era to Schneider's first-rank symptoms in order to focus on the key feature of schizophrenia, Maj 6 has previously concluded that the concept of autism, as revisited by Parnas 7, deserves attention. We believe that this element should be explicitly introduced in the current debate in the journal.