Here, we do not wish to repeat the arguments against the DSM definition, some of which are well known; they are based on examples that show that the definition is either too narrow or too wide. Neither, we do not want to go into the technical details of the functional definition. Instead, we would like to shortly discuss two broader points:
i) What is a 'scientific' definition of delusions and why do we need one?
ii) Why is a functional definition best suited?
The DSM is a manual used by both researchers and clinicians. For this reason, there is a multitude of constraints applying to the classification and the definition of mental disorders. 'Scientificity' or 'scientific adequacy' is only one of the possible criteria to be taken into account in order to define delusions. Others constraints include ease of applicability and reliability of daily diagnoses, political goals (e.g. securing that therapy costs are covered by insurances), social considerations (e.g. destigmatization), and therapeutic implications. All of these goals are equally important. However, finding a middle way between all these heterogeneous constraints comes with the risk of conflating different dimensions in the debate. For example, while it is socially and politically most significant to make a distinction between healthy and pathological, this dimension is of no big relevance for a scientific understanding of the mechanisms leading to this or that behaviour.
As philosophers of mind, we are foremost interested in the scientific understanding of the mechanisms that lead to a certain behaviour, be it or not considered a 'symptom'. One precondition for such an understanding is that one knows what one is trying to explain, namely, in this case, what delusions are. Thus, a clear definition of delusions is needed that should include only the scientific dimensions, but not include all dimensions included in the DSM, since it should be the basis for the scientific understanding. However, once we can investigate delusions empirically and conceptually based on a scientifically apt definition, we can learn about the underlying mechanisms leading to delusion. And this knowledge will most certainly have implications for therapy and is likely to have impact on the other dimensions as well (e.g., the scientific understanding of certain conditions like Down’s syndrome has greatly contributed to a destigmatization). Thus, without claiming that the scientific understanding of delusion is the most important part, we want to emphasize that a scientific is very useful and should be pursued. However, such a scientific understanding has to be based on scientifically apt definitions that do not include factors only relevant to other goals and dimensions.
For this reason, our definition might read a bit too simple at first sight. However – so we argue – it is scientifically apt. It specifies a phenomenon to be explained and calls for a scientific explanation. Moreover, the functional formulation of the definition also contributes to the goal of scientific aptness in different ways: First, the functional individuation of mental states is well known and considerably well established. Second, individuating mental states via their functional role (i.e. the role they play in relation to inputs, outputs, and other mental states) directly relates those states to behaviour: Not only is the empirical study of behaviour straightforwardly relevant for the definition of mental states; also, the identification of a mental state as being of this and that kind has immediate behavioural (and thus empirically testable) implications. Third, the functional definition is open to different possible explanations of the underlying mechanisms. In particular, it pairs very nicely with one of the most successful theoretical models, namely the two-factor-account: While one factor consists in an unusual experience that leads to the establishment of the delusional state in the first place, a second factor explains the immunization of this mental state towards counter-evidence and so explains why the state becomes delusional. In this way, the second factor is sufficient for delusionality, while the first factor grounds the subcategorization of delusions into subtypes (e.g. Capgras, delusion of persecution, delusion of control, etc.).