On Constructs and Mental Illness

Constructs are abstract concepts we use to organize available information for the purpose of (scientific) description and/or explanation. Constructs are ubiquitous in science in this sense; “gravity” and “temperature” are constructs, so are “intelligence” and “self-esteem”. Constructs can be immensely powerful when they capture features of the natural world (vs merely reflecting features of human interest), such as elements of the periodic table or fundamental particles in the standard model. But many scientific constructs do not map onto the world in such a powerful way as to “carve nature at its joints”. These constructs, while they do reflect features of the world (they are organizing information after all) also reflect human interests and goals to varying degrees.      

Diagnostic constructs in psychiatry are typically ways of organizing observed behaviors and reported experiences into particular categories or dimensions. To emphasize, for instance, DSM categories as “constructs”, is to emphasize the contingent nature of the organization and categorization. Major Depressive Disorder is one way of organizing observed behaviors and reported experiences, but we could adopt a different organization instead if we wanted to (“depressive neurosis”, “manic depressive illness”, “neurasthenia”, etc.)

Ideally, we want to “validate” our diagnostic constructs. That is, we want our organizations of behaviors and experiences to correspond with other “objective” things of interest, such as family history/genetics, neuroimaging, response to treatment, long-term outcomes, etc. If our constructs had such power, we might conveniently forget that they are constructs (we don't tend to think of “gravity” or “electron” as constructs). But in the absence of such validation (it would be more accurate to think of degrees of validation), to forget their contingent nature is to hinder our own scientific progress and to base our practical decisions on erroneous assumptions. To say that they are constructs is to remind ourselves of this contingency. It is in this sense that mental illnesses, i.e. specific diagnostic categories or dimensions, are “constructs”.

What about “mental illness” (the over-arching concept, not individual diagnoses or illnesses)? It too is a construct in the broad sense but compared to specific diagnoses such as Major Depressive Disorder, it is fuzzy and vague, residing more in ordinary language than in scientific theory. A lot depends on what we think the concept refers to. If someone sees “mental illness” as an explanatory concept – that by calling something a mental illness, we are hypothesizing the existence of a pathological process – they understand the concept very differently from someone who sees mental illness as a way of referring to problematic behaviors and experiences that have been identified as “abnormal” based on folk judgments (of proportionality, rationality, meaningful connections, etc.) and require professional care due to the associated distress, impairments, or harm. This means that it is less clear how we can think differently about the designatum of mental illness. Those who favor the explanatory view (erroneously in my opinion) would propose a different explanation, but given that I think the explanatory view is incorrect to begin with, these alternatives explanations don’t fare any better. I think the real question often boils down to: when someone says that mental illness is a construct that can be replaced by a different construct, are they referring to a different designatum, or are they referring to the same designatum (behaviors identified by folk judgments as abnormal, requiring some form of professional care)? If the designatum is the same, is this merely a preference to use a different term for the designatum because of perceived connotations of the relevant terms? A lot of the time I think the debate is merely terminological but gives the illusion of being about something deeper.