DRAWING THE LINE

When does a psychological experience qualify as pathological? Is it hopeless to look for an objective line between ‘normal’ and ‘pathological’ when it comes to our mental life? After all, aren’t we all ‘a bit crazy’?

The issue of determining what counts as pathological is obviously critical for psychiatric research and practice. One might even say that one of the main goals of psychiatry lies in identifying as accurately as possible who is entitled to medical treatment. Drawing such a boundary also has important ethical and clinical implications, as receiving a psychiatric diagnosis can significantly affect one’s life. Yet, researchers, practitioners, and individuals diagnosed with a mental condition disagree about whether psychiatric categories should be seen as harmful, dehumanizing, or liberating. For some, such categories inevitably perpetuate stigma and discrimination (Haslam [2011]; Zachar [2015]); for others, psychiatric diagnoses denote an essential aspect of one’s identity, whereas claims such as ‘we are all a bit x’ (where x stands for any psychiatric condition) are misguided and harmful (Botha et al. [2020]).

Psychiatry is therefore caught in a bind. As a medical discipline, it should correctly identify people in need of treatment and grant them access to appropriate support and services. At the same time, it should be wary of wrongfully pathologizing behaviours that could be understood as ‘atypical’ manifestations of non-pathological phenomena.

Dimensional approaches to psychopathology, according to which symptoms vary continuously from health to disease along one (or few) dimensions, offer a compelling way out of the bind. They do justice to the continuity of human psychological variation, as they understand the difference between health and pathology as a matter of degree. To better visualize this, think about a spectrum of colours fading into one another. This idea is based on the observation that behaviours typically associated with mental conditions are continuously distributed in the general population. For instance, there is now robust evidence showing that experiences once associated exclusively with psychosis—such as hearing voices—are in fact rather common (Johns and van Os [2001]). The dimensional view is also supported by genetic data (Jang [2005]; Knopik et al. [2017]). In this sense, dimensional accounts encourage the idea that there is no principled way to set the threshold between normal and pathological experiences.

At the same time, both clinical practitioners and researchers are aware that such a threshold often needs to be established for practical purposes. Thresholds are usually conceptualized in two ways. Clinical practitioners—who tend to be less interested in the nature of the health–pathology distinction—set the cut-off point merely to serve pragmatic purposes, but do so without embracing any strong commitment regarding the existence of a threshold (or the lack thereof). Clinical and theoretical researchers, by contrast, often take the distinction between health and pathology to be conventional. Sigmund Freud ([1938], p. 81) represents a paradigmatic example of such a view: ‘It is not scientifically feasible to draw a line of demarcation between what is psychically normal and abnormal; so that distinction, in spite of all its practical importance, possesses only a conventional value’. As opposed to the pragmatic considerations mentioned above, the conventional view takes a much firmer stance: the health–pathology distinction is believed to be conventional because of an actual vagueness between the two. Think again about the spectrum of colours: if there really is no way of telling where blue ends and green begins, then a dimensional model is what better captures the underlying phenomenon of continuity. Similarly, if health and pathology are truly continuous, then we cannot but set thresholds conventionally while knowing—deep down—that there are ‘no joints to be found’ (Kendell [1975]).

Despite their popularity, dimensional approaches leave some important problems unsolved. One major issue concerns how to draw the line between health and pathology. Even if we see thresholds as pragmatic or conventional, we still need some non-arbitrary criteria to determine where to set them. Several clinical thresholds are determined by considering quantifiable factors, such as duration of symptoms—for example, at least fourteen consecutive days for major depression. Yet, one may legitimately wonder whether there is any actual difference between two individuals who experience a given symptom for thirteen versus fourteen days. Another issue is that dimensional models typically draw an unwarranted inference from the continuity observed at the behavioural level to a quantitative difference between health and pathology at the genetic level.

The proposal we develop in our article resists such an inference and provides a characterization of mental disorders as complex states that are qualitatively different from healthy ones. This approach sheds light on a variety of issues.

The first tenet of our model is that mental disorders are best understood as highly heterogeneous constellations of psychological and behavioural traits along multiple dimensions. This emphasizes that individuals are all very different when it comes to identifying the specific biological, environmental, and social factors underpinning their disordered manifestations. We maintain that these profound individual differences cannot be fully grasped by the conventional-threshold view insofar as this takes normality and pathology as extremes of a unidimensional continuum.

The second tenet of our model is that thresholds should be conceived as ontological, rather than conventional, meaning that they do not just depend on how we construe them, but rather on objective changes in a biological system. To give substance to this idea, we make use of the conceptual architecture provided by Conrad Waddington ([2008]). Waddington’s epigenetic landscape depicts biological development as a process where a ball rolls down a sequence of valleys, each representing an alternative developmental path.

By analogy, mental conditions—be they healthy or pathological—can be conceptualized as endpoints of alternative developmental trajectories (A, B, C, and D in figure 1), which are generated by bifurcations where an individual can take one path or another. Ontological thresholds, in our view, are nothing but bifurcations of this sort. An individual crossing such a threshold would typically enter a qualitatively different state, and a different constellation of behavioural and psychological features would emerge. This is exactly what a psychiatric diagnosis attempts to capture: the moment where a person shifts from a vulnerable state to a pathological one.

The third tenet of our model is the emphasis that we place on the time variable. Representing mental disorders within an epigenetic framework allows us to describe psychological development as a stable flow of transitions from one constellation to another. On this view, constellations are both stable and dynamic. They are stable because psychological development can become canalized over time within specific trajectories, as in the case of neurodevelopmental conditions such as autism, which significantly constrain future development. Yet, psychological development is also flexible, which means that constellations develop and change in response to interactions among an individual’s characteristics and environment. This is why, in many cases, recovery from mental disorders is possible: life events, including treatment, may exercise a key role in pushing an individual across the threshold(s) in one direction or another.

Conceptualizing mental disorders as constellations that develop over time in an epigenetic framework has major implications for psychiatric intervention and treatment, particularly in terms of tracking individual trajectories and paying attention to relevant transitions across psychological states.

Some of these transitions concern precisely the threshold between healthy and pathological states. Vulnerability, for instance, may be understood as a paradigmatic ‘brink state’, where an individual gets closer to a pathological state and is therefore more likely to gravitate towards it, depending on the circumstances. While some individuals go from being vulnerable to developing a psychiatric condition, others revert to a condition of health—they cross the threshold from vulnerability to a healthy state.

This relates to another relevant transition discussed in our model, namely, recovery. Since development follows the time arrow, the range of developmental potentials narrows down over time—in other words, one never truly goes back from the present state to a previous one. Recovery should therefore be conceptualized as the generation of novel constellations, rather than as the return to previous healthy states. This does justice to a powerful observation from clinical practice: if an individual effortfully achieves health after having experienced a disordered condition, their constellation and developmental potential will differ from that exhibited by a person who has never fallen ill in the first place (for instance, they may be at higher risk of relapse). After all, generating and maintaining a novel qualitative state requires bringing about significant and stable changes, and interventions may not have unlimited power in shaping the future path.

An important note, before concluding: Seeing mental disorders as qualitatively different from healthy states does not imply that they fall outside the range of typical human variation. On the contrary, mental disorders are dynamic and individualized states that cannot be reduced to empty and potentially dehumanizing labels. Our model does justice to the dynamic nature of psychological development and to the influence that life events exercise on health and pathology by modifying individual paths in drastic ways.