In his editorial on the interpersonal domain, Hobson (Reference Hobson2003) asserts that the analysis of intersubjective engagement in the therapeutic dyad is essential to the understanding of subjective meanings and their role in the manifestation of psychiatric disorder. In the same issue of the Journal, Lanman et al (Reference Lanman, Grier and Evans2003) describe their attempts to determine a measure of ‘fit’ between individuals in a ‘couple system’.

Both these papers acknowledge a fundamental fact concerning all human relationships; that they are, in their totality, the interaction between one personality and another. What is striking, however, both in these papers and in other recent literature concerning, in particular, personality disorders (Reference Tyrer, Duggan and CoidTyrer et al, 2002), is the lack of any discussion concerning the specific role the therapist's/clinician's personality plays in shaping the therapeutic relationship.

Hobson's use of Donne's metaphor (‘No man is an Island, entire of it self’) captures what I believe to be the sine qua non of personality disorder; namely, that personality disorders can only be understood in the context of interactions between personalities; that the construct of personality disorder cannot exist in isolation. This notion is akin to the distinction made between ‘primary’ and ‘secondary’ qualities by the philosopher John Locke. In a psychiatric context one might consider schizophrenia to be a primary phenomenon, an integral part of the individual, whereas personality disorder, being contingent on an interaction with another, is secondary.

If one can accept the notion of personality disorder as a consequence of the interaction between two personalities, then surely it behoves members of the psychiatric profession to consider how their personalities influence the therapeutic relationships that lie at the heart of the discipline. That this appears, historically, not to have been the case is revealed by Lewis & Appleby's (Reference Lewis and Appleby1988) seminal paper. While amply demonstrating psychiatrists' negative attitude towards individuals with personality disorder, the authors failed to address the possibility that this might be a function, in part, of the psychiatrists' personalities.

If we are to be ‘scientific’ about studying interpersonal functioning, then perhaps the first step might be to consider a systematic evaluation of both personalities involved in the therapeutic dyad. One possible method might employ a dimensional assessment of personality that would, in turn, help define how different personalities ‘fit’ together. For example it might be reasonable to expect a clinician, scoring highly on the ‘openness’ dimension of the NEO–PI–R (Reference Costa and McCraeCosta & McCrae, 1992) to fit well with a patient scoring much lower on the same scale.

If this were shown to be the case, it could have important ramifications for resource allocation, both in psychotherapy and in the wider psychiatric field, allowing individual personalities to be fitted together in order to better facilitate the therapeutic relationship. An appreciation of the role their own personalities play in the construct known as personality disorder, might also diminish psychiatrists' negative attitudes to the disorder they appear to dislike.