The clear definition of primary delusions helpfully provided by Owen et al (Reference Owen, Harland and Antonova2004) includes both that one comes to believe new things (change in meaning) and arrives at such beliefs in a new way (transformation of experience). This requirement for a transformation of experience seems to me to require an abnormal mental mechanism that is permitted in a rationalist account of delusions. The distinction between primary and secondary delusions is one that eludes many authors (e.g. Reference Hales, Yudofsky and TalbottHales et al, 1999: pp. 432–434) in addition to myself and Professors van Os and Delespaul. We would differ I suppose in that while they view all delusions as secondary, I think those that do not share the mechanism of primary delusions are not really delusions at all.

Wernicke's work on aphasia suggested that mental functions were localised, but this is quite a separate issue from whether their mechanism is modular. This is evident in Wernicke's description of mental pathology as a ‘loosening up of the firm network of association’ (Reference JaspersJaspers, 1963: p. 536). Such an empiricist account of mental pathology is surely incompatible with modularity as proposed by the rationalist philosopher Fodor.