Dr Ambelas raises the important relationship between the premorbid characteristics of individuals who later develop schizophrenic illnesses and the syndrome first described by Hans Asperger as ‘autistic psychopathy’ in childhood (Reference Asperger and FrithAsperger, 1944). Asperger related his clinical picture to Bleuler's concept of autism in schizophrenia and wrote that, ‘All but the last mentioned feature (dereistic thinking) can be found in the type of personality disorder to be described here’. But ‘While the schizophrenic patient seems to show progressive loss of contact, the children we are discussing lack contact from the start’. Investigating this association, Tantam (Reference Tantam1988) found that 18 (21%) of 86 people with Asperger syndrome later developed some form of psychosis.

The status of Asperger disorder/syndrome (DSM—IV (American Psychiatric Association, 1994) and ICD—10 (World Health Organization, 1992)) within the class of autistic spectrum or pervasive developmental disorders (DSM—IV) has been much debated. These disorders are characterised by delays or deficits in social relatedness, reciprocation, and understanding social interactions. The term pervasive developmental disorders was first introduced in DSM—III (American Psychiatric Association, 1980), with Asperger disorder only separated from other pervasive developmental disorders in DSM—IV. Pervasive developmental disorders not otherwise specified constituted the majority of cases in the DSM—IV field trials. Further subdivisions of pervasive developmental disorders are likely in revisions of DSM resulting from empirical evidence and consensus of opinion. Thus, Ambelas's target of a ‘primary candidate’ at this stage might be the broader class of pervasive developmental disorders, excluding autism, rather than Asperger syndrome per se.

Cohort studies such as the National Child Development Study (NCDS) cast some light on the issue. The epidemiology is arguably similar, with S+ schizophrenia having a lifetime prevalence of 8 per 10 000, and in the NCDS at age 7 the gender split was 20:13 (i.e. 1.5:1). While Ehlers & Gillberg (Reference Ehlers and Gillberg1993) using their own criteria estimated a minimum prevalence of 3.6 per 1000 children (7-16 years of age) and a male to female ratio of 4:1, using more liberal criteria their prevalence was 7 per 1000, with a gender split of 2.3:1.

Most authors agree with Tantam that the core of Asperger syndrome consists of disabilities in communication, socialisation and non-verbal expression, with conspicuous clumsiness and special interests. Cohort studies suggest that there are indeed deficits in at least some of these areas in children who go on to develop schizophrenia in adulthood. In the NCDS, we found these children more often rated as over-anxious and hostile in their relationships with adults and other children, and this was both more marked and present earlier in boys (Reference Done, Crow and JohnstoneDone et al, 1994). At ages 7, 11 and 16, their teachers noted the children were mispronouncing words more often than the rest of the cohort. At 11, there were increased rates of speech difficulties, and at 16 they were poor on English ability. There are therefore difficulties in communication, although it is not clear that these are comparable to the ‘odd, pedantic, stereotypic speech’ that is described in Asperger syndrome. Interestingly, at each age they were delayed in reading ability, although such deficits are not recorded as characteristic of Asperger syndrome. At age 11, girls but not boys among those who later developed schizophrenia were rated as withdrawn (i.e. distant, cut-off from people, avoiding communication), evidence perhaps of difficulties in non-verbal communication. However, at age 7 the girls in all respects manifested normal social behaviour, suggesting that girls who, in adulthood, develop schizophrenia might display a characteristic developmental trajectory (i.e. a decline in social relatedness and reciprocation between childhood and adolescence).

Perhaps the most interesting parallel is the one to which Ambelas draws attention, between the increase in neurological soft signs that we have observed and the clumsiness and stereotypy of movement that is described in Asperger syndrome — a clue to the neurological basis or bases of the two clinical pictures. At age 7, the children who, in adulthood, developed schizophrenia were more likely to be rated as having difficulties in coordination, and at age 16 were more likely to be described as clumsy. Delays in learning to stand, walk or speak are linearly related to later risk of schizophrenia (Reference Isohanni, Jones and MoilanenIsohanni et al, 2001) as also was delay in potty-training, a finding that corresponds to an increased incidence of failures of continence observed in the NCDS cohort.

Each of these findings suggests some commonality between features of pervasive developmental disorders (including Asperger syndrome) and those that precede schizophrenia, notwithstanding obvious differences in methods of data collection. There remains the question of time course to which Asperger drew attention — the features of Asperger syndrome are present early, whereas those of schizophrenia show an element of progression. This difference is exemplified by a progressive decrease in language scores preceding the onset of schizophrenia (Reference Fuller, Nopoulos and O'LearyFuller et al, 2002).

The key question is why does this constellation of features come together? One finding from the NCDS sample points to the neurological substrate. At age 7, children who, in adulthood, develop schizophrenia were more likely to be rated as ambidextrous for handwriting by their mothers and were less lateralised on a square ticking task at age 11. There is evidence that lateralisation is a major determinant of the acquisition of words as well as of other aspects of cognitive ability (Reference Crow, Crow and DoneCrow et al, 1998). We propose that development of hemispheric dominance for the components of language is relevant to the similarities of the conditions described by the term Asperger syndrome and those that precede the onset of schizophrenic psychoses, as well as to differences in their time course. Thus, language — the core characteristic of the species — with its context in social interaction and its matrix in the lateralisation to the two hemispheres, is the function that varies between individuals and accounts for the similarities between the early developmental anomalies that were identified by Asperger and those that can now be seen as precursors to psychotic illness.