Introduction: Poor mentalising has been linked to particular psychotic symptoms (e.g., paranoia) and attributed to selective disruption of a mentalising module. Mentalising, executive planning, and disengagement were tested in patients with schizophrenia and healthy controls in order to evaluate this view against non-modular accounts that attribute poor mentalising to generalised difficulty entertaining any hypothetical state of affairs and/or difficulty inhibiting salient misleading information. Method. Mentalising and disengagement were tested in a picture-sequencing task using false-belief and capture stories - the former require inferences of mental states; the latter test ability to inhibit salient misleading cues. Executive planning was tested using the Tower of London task. Results. Whereas patients as a whole showed impairments of mentalising, executive planning and disengagement, false-belief picture sequencing ability significantly predicted the odds of being a patient, after adjusting for all other task measures. No evidence was found linking poor mentalising to positive symptoms. Conclusions. Our findings support the existence of a mentalising module, which is selectively disrupted in some patients with schizophrenia. Null results concerning the links between poor mentalising and positive symptoms are discussed in relation to current views on whether poor mentalising is best conceptualised as a state or trait marker of psychosis.