Abstract
Background:
Empirical explanation and treatment repeatedly fail for psychiatric diagnoses. Diagnosis is mired in conceptual confusion that is illuminated by Ludwig Wittgenstein’s later critique of philosophy (Philosophical Investigations). This paper examines conceptual confusions in the foundation of psychiatric diagnosis from some of Wittgenstein’s important critical viewpoints.
Argument:
Diagnostic terms are words whose meanings are given by usages not definitions. Diagnoses, by Wittgenstein’s analogy with ‘games’, have various and evolving usages that are connected by family relationships, and no essence or core phenomenon connects them. Their usages will change according to the demands and contexts in which they are employed. Diagnoses, like many psychological terms, such as ‘reading’ or ‘understanding’, are concepts that refer not to fixed behavioural or mental states but to complex apprehensions of the relationship of a variety of behavioural phenomena with the world. A diagnosis is a sort of concept that cannot be located in or explained by a mental process.
Conclusion:
A diagnosis is an exercise in language and its usage changes according to the context and the needs it addresses. Diagnoses have important uses but they are irreducibly heterogeneous and cannot be identified with or connected to particular mental processes or even with a unity of phenomena that can be addressed empirically. This makes understandable not only the repeated failure of empirical science to replicate or illuminate genetic, neurophysiologic, psychic or social processes underlying diagnoses but also the emptiness of a succession of explanatory theories and treatment effects that cannot be repeated or stubbornly regress to the mean. Attempts to fix the meanings of diagnoses to allow empirical explanation will and should fail as there is no foundation on which a fixed meaning can be built and it can only be done at the cost of the relevance and usefulness of diagnosis.
Introduction
In a well-known but enigmatic paragraph at the end of his posthumously published Philosophical Investigations, Ludwig Wittgenstein (2001) says:
The confusion and barrenness of psychology is not to be explained by calling it a ‘young’ science … in psychology there are experimental methods and conceptual confusion. The existence of the experimental method makes us think we have the means of solving the problems which trouble us; though problem and method pass each other by (Wittgenstein, 2001: Part II, sec xiv, p.197e).
The comment points back into this dense and exacting work and the present paper will look at some of the arguments.
Wittgenstein was talking of psychology rather than psychiatry and was writing before 1950, but this statement should resonate with many clinicians’ experience that diagnoses neither predict the phenomena and outcomes seen in practice nor relate well with contradictory research findings. In particular, diagnoses rarely seem to explain the patients’ problems (Rosenman et al., 2003). This seems a practical problem to be solved by science, which, among other things, will ‘… discover neurobiological endophenotypes bridging the gap between behavioral phenotype and genotype’ (Hasler et al., 2004). From Wittgenstein we may say that it is not a practical problem of incomplete scientific knowledge. It is a philosophical problem of concepts. The problem will not be solved by maturation of a young science’s empirical techniques.
The philosophy of psychology is an aspect of all Wittgenstein’s work. Here we will examine three arguments that expose a central misconception that clinical psychiatric diagnoses point to particular mental processes. These arguments:
word meanings are given by their usages and the problems of definition;
family resemblance and essence;
the misidentification of mental processes;
are among the many multi-faceted arguments in Philosophical Investigations (Wittgenstein, 2001). We trace their implications for diagnosis in psychiatry. They are only a few of the important ideas in his later writing. We quote extensively from Philosophical Investigations to introduce Wittgenstein’s writing, which has not been widely applied to concepts in psychiatry.
Word meanings given by usages
A diagnosis such as ‘depressive disorder’, as used by psychiatrists, is a term in a diagnostic language. The DSM (American Psychiatric Association, 2000) and ICD (WHO, 1993) attempt to give clear and restrictive definitions – to firmly tie labels to the behaviours, appearances and utterances peculiar to the disease. Obviously, a term should have a clear and unambiguous meaning so that we know exactly to what the term refers. But, says Wittgenstein, this idea of reference is a very primitive view of what we do with language (Wittgenstein, 2001, para 2). Although we use the term ‘major depressive disorder’ in the way the DSM directs, we do all sorts of things with this term. ‘There are … countless different kinds of use of what we call “symbols”, “words”, “sentences” (Wittgenstein, 2001, para 23). The profession uses the term not only to decide treatment but also to select research subjects, to persuade or inform the public, to argue in court, to organise services, and so on. Each time the term is used, its usage extends in ways that cannot be foreseen in any attempt to regulate the language.
Is then the DSM definition correct? YES, but only for a very circumscribed region of the common professional usages. We can make it is a complete definition only if we exclude many current and future professional usages (Wittgenstein, 2001, para 3).
A term such as depressive disorder is established within the purposes and contexts in which it is used: ‘[the] language and actions into which it is woven: a language game’ (Wittgenstein, 2001, para 7). The profession has many different but legitimate language games, which include treatment, explanation, law, research, and so on. If you do not keep the different games in mind, you will keep trying to ask the question ‘What is depression?’ (Wittgenstein, 2001, para 24) and think that you can use empirical methods to find the answer. If you can restrict the usage of the diagnostic term to one game, you also restrict its usefulness and stop usage expanding to cover the widening activities the profession is called on to deal with.
It seems unlikely that the profession will stop using the term in all sorts of ways. There is not even a rough consensus, let alone agreement, on what constitutes depression (Kirk and Kutchins, 1992; Parker, 2008). Notwithstanding the DSM, there will almost certainly never be a final and agreed definition of the condition. A pluralistic and outward-looking discipline will stretch and modify its language and concepts continuously as it grapples with its problems (Berrios, 1999).
The diagnosis can be understood only as it is used within the contexts or activities that give it meaning. The term is understood by looking at the activities of the profession that uses it, NOT by reference to an object or specific mental process.
Definition and naming
The apparent clarity of the definitions in the DSM and ICD misleads us. The definition seems clear, and appears to relate to some thing in the world. Is the definition a fiction? No, the DSM is built from descriptions of phenomena to give a definite and distinct definition of what we should call depression. However, if you examine the descriptions incorporated in the definition you find that many of them call out for definition or interpretation. The usage of words such as ‘depressed mood’, ‘significantly’, ‘inappropriate’, ‘agitation’ (DSM criteria for major depressive disorder (American Psychiatric Association, 1994)) also depends on the many language games in which they are used. Naming them does not finalise them, but simply points to a need for further description. Anyone who has tried to teach or to turn diagnostic criteria into diagnostic instruments will understand that this is not trivial. We have to examine the contexts in which the elements of the definition are used. Rather than clearly defining something that is produced by a mental process, the definition may do no more than to move the uncertainty to the next fork in the path:
[A] sign … may be sometimes a word and sometimes a proposition … [and this] depends on the situation in which it is uttered or written. … naming and describing do not stand on the same level: Naming is a preparation for description. … Naming is so far not a move in the language-game any more than putting a piece in its place on the board is a move in chess. (Wittgenstein, 2001, para 49).
Essence and family resemblance
We use a term such as ‘depressive disorder’ flexibly and intricately in a range of cases. We assume that, because we use the term across these cases, there must be something common or essential underlying them. Discovering that essence seems to be the task of psychiatry and that essence appears necessary so that we can set rules for treatment as well as for experiment and research.
But this assumption of an essence is ‘a craving for generality … the tendency to look for something in common to all the entities which we commonly subsume under a general term’ (Wittgenstein, 1958). Cases of ‘depression’ are related not by an essence but by the extensions of meaning that grow as a term is used to grapple with evolving practice in the world. Wittgenstein gives an analogy with the concept of games:
Consider for example the proceedings that we call ‘games’. I mean board-games, card-games, ball-games, Olympic games, and so on. What is common to them all?—Don’t say: “There must be something common, or they would not be called ‘games’”—but look and see whether there is anything common to all.—For if you look at them you will not see something that is common to all, but similarities, relationships, and a whole series of them at that. To repeat: don’t think, but look! (Wittgenstein, 2001, para 66).
We have no trouble identifying a game, but what is the essence that makes a game a ‘game’? There is none. As soon as you offer an essential feature that makes a game a game, there is a game that will not fit. The use of the term expands as new activities develop or new games are invented. Usage spreads by metaphor or by idiosyncratic resemblances as we may talk legitimately about the ‘mating game’ or may say ‘he is just playing games with you’. Games form a family of activities that resemble each other in family ways (‘… the various resemblances between members of a family: build, features, colour of eyes, gait, temperament, etc. overlap and criss-cross in the same way’ (Wittgenstein, 2001, para 67)). There is no sharp boundary for the concept:
… the concept ‘game’ is a concept with blurred edges.—“But is a blurred concept a concept at all?”—Is an indistinct photograph a picture of a person at all? Is it even always an advantage to replace an indistinct picture by a sharp one? Isn’t the indistinct one often exactly what we need? (Wittgenstein, 2001, para 71).
For clinical purposes, a loose inclusive idea may be more useful than a sharp exclusive one.
No essence underlies all the cases that we include in the concept no matter how we try to specify the core or pathognomonic features or try to exclude the untypical or irrelevant cases: ‘In order to find the real artichoke, we divested it of its leaves. … what is essential …, however, was not hidden below the surface of this case, but this “surface” was one case out of a family of cases …’ (Wittgenstein, 2001, para 164). The cases we diagnose with ‘depression’ join a family. Each case extends the family and no single or essential characteristic is found in each.
Wittgenstein returns repeatedly to family resemblances and families of cases and he gives examples that are peculiarly relevant to ideas of mental states and processes. Take, for example, ‘reading’, which we think of as a mental process. How do we discriminate actual ‘reading’ from various activities which we might call ‘proto-reading’ (as a child learning to read might do on the way to becoming a reader), or reading music, or reading a comic or reading facial expression? We see them all as ‘reading’. They all require mental processes, but there is no reason to believe that it will be the same process or set of processes in each case. We know already that many human behaviours that are related conceptually need not share mental processes; words in songs are encoded differently from words in prose (Platel et al., 2003). What we call ‘reading’ is a family of mental activities that resemble each other and no sharp boundary distinguishes ‘reading’ from similar activities. We cannot find a single criterion, or a single process of ‘reading’.
Even though phenomena may not be connected by a physical process, they can hang together in a concept. Wittgenstein gives the metaphor of a thread: ‘… the strength of the thread does not reside in the fact that some one fibre runs through its whole length, but in the overlapping of many fibres’ (Wittgenstein, 2001, para 67).
For depression, no single criterion or small number of criteria and no single mental process or small number of mental processes runs through what we unhesitatingly identify as depression.
Mental processes
The most subtle of these arguments is exposed in Wittgenstein’s discussion of understanding as a mental process: ‘try not to think of understanding as a “mental process” at all. – For that is the expression that confuses you. … In the sense in which there are processes (including mental processes) which are characteristic of understanding, understanding is not a mental process’ (Wittgenstein, 2001, para 154).
We are apt to think that ‘understanding’ is a process, perhaps analogous with the wave of neuronal activation that goes with the movement of an arm – but this analogy leads us astray. Understanding is something that we impute to various actions we observe in others or feelings in ourselves. We say I understand when I say ‘Ah ha!’ to myself. We say someone understands when she solves an arithmetic problem. I will say you understand when you laugh at a joke or translate a passage from a foreign language.
If you look in the brain for the evidence of the mental process of ‘understanding’ you will find evidence of mental processes; in the parietal lobe while calculating, speech areas while translating. But understanding cannot be found there because it is not there. The concept of ‘understanding’ is a complex network of usages by an observer (or self-observer) that refer to this family of behaviours. In some cases the understanding is the ability to continue with a task, in others it is the ability to respond or not respond. These are all judged as ‘understanding’ by reference to the activities of the human in the world. These activities are all ‘understanding’ but those ‘understandings’ are not reflected in the brain as a single mental or brain process. The ‘understanding’ is not a process in the person but a judgement of the actions of the person made by an observer or self-observer.
All these arguments apply to what is called ‘knowing’ (knowing how to swim, knowing the way home, knowing the words of a song, knowing myself), and they can be applied directly to diagnoses such as depression. In the brains of the people called ‘depressed’ there are numerous mental processes. In some there is autonomic arousal, in others, processes associated with sleep and wakefulness. There will be brain processes associated with appetite and satiety. The diagnosis flows from the identification of phenomena that are readily visible and not from the identification of a hidden process. There will not be a brain process that is ‘depression’ because the diagnosis does not reside there. Depression can be seen as analogous with ‘understanding’, of which Wittgenstein says:
We are trying to get hold of the mental process of understanding which seems to be hidden behind these coarser and therefore more readily visible accompaniments. But we do not succeed; or, rather, it does not get as far as a real attempt. For even supposing I had found something that happened in all those cases of understanding, – why should it be the understanding? (Wittgenstein, 2001, para 153)
Many diagnostic phenomena reflect mental processes in the person but the phenomena are members of the family of depressive signs because we see their relationship with other phenomena and with the world. The mental processes in waking early in grief may not be different from waking early in expectation. It is a cardinal sign of depression not simply by the particularities of the sleep disturbance but because it is related with other phenomena, such as agitation. The particularities that make it part of depression cannot be found by looking at the processes in a laboratory.
Summary
There are several problems with many psychiatric diagnoses exemplified by the clinical diagnosis of depression (or major depressive disorder in its codified form).
Firstly we know the meaning of the term only by its usage. Even within the expert tribe the term is used in many different ways and any standard definition will cover only a small area of its accepted usage. It cannot be a suitable target for empirical brain research except by the most restricted defining that ignores the actual professional use of the term.
Second, the definition that appears so solid in the DSM turns out to move the definition problem one step further. The elements of the definition turn out to require description themselves.
Third, the attempt to find the essence of depression, the attempt to get to the endophenotype and its mental process, looks past the family of cases that constitute the diagnosis. The diagnosis has no essence, no endophenotype and no specific underlying process that can form the target of an empirical investigation.
Fourth, the concept of depression appears to be composed from a large number of observed states and, although we do not hesitate to identify them as depression, they are not identified by fitting criteria. They are identified in the same way that we recognise family membership or identify a game as a game or a joke as a joke: a family of states and activities that need not have any single or even a small number of processes in common.
These arguments are aspects of closely related arguments which include the idea of families of meaning expressed by families of usages of the word.
The argument from understanding raises the identification of processes and how we locate them. Wittgenstein does
Implications
Implications for diagnosis
Diagnostic terms such as ‘depressive disorder’ are used in different ways according to context. They may not be properly understood outside the situation in which they are being used. Contexts and situations vary even within the profession, so that the terms do not communicate reliably unless the phenomena and the contexts are communicated along with the diagnosis. Diagnosis without context information is more likely to mislead than to provide useful information. Patients’ conditions will need to be described case by case.
If there is no essence (or essential phenomenon) in the family of cases that we identify as depression, there can be no monothetic definition of the condition (that is a definition composed of an essential element or criterion). A polythetic definition will be a list of phenomena that have been identified in cases of depression now and in the past. Different phenomena have been considered part of depression in the past (Berrios, 1988) and will be different again in the future. What use then can diagnosis be?
Diagnoses are not pointers to an underlying process but ‘… rather are set up as objects of comparison which are meant to throw light on the facts … by way not only of similarities, but also of dissimilarities’ (Wittgenstein, 2001, para 130). The diagnosis may illuminate a case by how the case fits but also by how it does not fit; how it resembles but also how it differs from some paradigmatic model cases embodied in the diagnoses: ‘… we can avoid ineptness or emptiness in our assertions only by presenting the model as what it is, as an object of comparison—as, so to speak, a measuring-rod; not as a preconceived idea to which reality must correspond. (The dogmatism into which we fall so easily …) (Wittgenstein, 2001, para 131).
Diagnoses have important uses. ‘If I tell someone “Stand roughly here” – may not this explanation work perfectly? … But isn’t it an inexact explanation? … Only let us understand what “inexact” means. For it does not mean “unusable”’ (Wittgenstein, 2001, para 88). They point usefully but loosely to the domains in which phenomena and their relationships are found and shine a light on phenomena that otherwise may not be seen by the inexperienced eye. Diagnosis will be useful to the patient to know what services to approach and to the clinician to have a starting point for enquiry and know where useful scientific knowledge may be found. Diagnoses are deeply embedded in the social legitimation and demarcation of roles for the profession and community. But there are clinical and scientific purposes for which diagnosis cannot be used.
Implications for clinical and scientific explanation
Diagnosis gains professional and popular potency because it seems to ‘explain’ the problems. The need to know hidden explanatory mechanisms infects the approach to the description of phenomena. A description of an experience, ‘I feel depressed in this way’, becomes an explanation, ‘I feel this way because I have depression’. The very thing we wanted to explain, the apparent ordered collection of phenomena, becomes the explanation of the phenomena: ‘Our mistake is to look for an explanation where we ought to look at what happens as a proto-phenomenon. That is, where we ought to have said: this language-game is played’ (Wittgenstein, 2001, para 654).
As an explanatory model, diagnosis is empty but belief in an explanatory process introduces a conceptual confusion into the very foundation of thinking:
How does the philosophical problem about mental processes and states … arise? The first step is the one that altogether escapes notice. We talk of processes and states and leave their nature undecided. Sometime perhaps we shall know more about them. But that is just what commits us to a particular way of looking at the matter. For we have a definite concept of what it means to know a process better. (The decisive movement in the conjuring trick has been made, and it was the very one we thought quite innocent.) … (Wittgenstein, 2001, para 308)
Before we start on our problems we already have a commitment to a particular way in which the problems will be solved.
This confusion is introduced right at the beginning of training in psychiatry and spreads through wider popular understanding. It is a confusion that cannot be elucidated by increasing sophistication of many empirical techniques because the techniques assume the underlying process. This is where ‘problem and method pass each other by’ and it explains why research based on a diagnosis so often founders.
The DSM from 1980 departed from previous nomenclature, styling itself as ‘a descriptive approach that attempted to be neutral with respect to theories of etiology’ (American Psychiatric Association, 2000). It directly translated the Research Diagnostic Criteria (RDC) (Spitzer et al., 1975, 1978), which had created a rudimentary but tightly controlled language that attempted to carve out phenomenologically coherent fragments of psychopathology. The RDC did not pretend to describe the whole of any domain and did not explicitly assume a mental process underlying each fragment. The translation to the DSM appears to have lost sight that the RDC was a description in language of a fragment of co-occurring phenomena that applied to a subset of subjects. It was not an inclusive description of the domain. The RDC diagnoses were orderings of phenomena that were a preparation for the then contemporary empirical research. It was not a label for a mental process or an explanatory mechanism. To talk about diagnosis was to talk ‘… as we do about the pieces in chess when we are stating the rules of the game, not describing their physical properties. The question “What is a word really?” is analogous to “What is a piece in chess?” (Wittgenstein, 2001, para 108).
The confusion applies to physical, psychological, evolutionary and social explanation (Breen et al., 2011), as well as the unresolvable wrangling over the nature, cause and treatment of conditions such as depression (Parker and Hickie, 2007). It is not that the explanations are inadequate but that the object of the explanations is a mobile and variably used word, not a mental process. That is not to deny the existence or relevance of mental processes, only to recognise that a diagnosis is not a pointer to a particular mental process that is the disease. This argument appears ‘… to deny the yet uncomprehended process in the yet unexplored medium. And now it looks as if we had denied mental processes. And naturally we do not want to deny them’ (Wittgenstein, 2001, para 308) but it is the identification of confusion, not a subscription to ‘behaviourism’ or a ‘mind-brain’ problem.
Thus many scientific investigations cannot be done on a diagnosis. These include the natural history, the causes and the neurophysiology of depression. Such studies will always produce empirical ‘findings’, but those findings will vary according to the diagnostic usages and the contexts that are embedded in the selection of subjects. Alternatively the findings will be diffused to meaninglessness by the phenomenological heterogeneity of the subjects. Scientific studies may be done on some (not all) individual phenomena but that will not tell us what makes them part of depression. (This point may be understood by analogy with ‘games’. For example, sports physiology research such as fast twitch muscle in sprinters (Jansson et al., 1990) informs us about certain aspects of a particular game but it tells us nothing about the concept of games or even of athletics.)
Implications for practice
The belief that a specific mental process underlies a diagnosis dictates many of the clinical approaches to problems.
This picture of a specific mental process calling for specific treatment seems to grip psychiatry and embeds an explanatory model where explanation may not be possible. Wittgenstein uses the term ‘picture’ (Bild) to describe both representations and conceptions (Baker, 2001). There is a particular picture of depression embedded in the training of psychiatrists which forms the professional language and regulates their perception of the problem: ‘We predicate of the thing what lies in the method of representing it … we think we are perceiving a state of affairs of the highest generality’ (Wittgenstein, 2001, para 104). It directs which phenomena are seen and how the concept is to be constructed: ‘A picture held us captive. And we could not get outside it, for it lay in our language and language seemed to repeat it to us inexorably’ (Wittgenstein, 2001, para 115). We seem to have no other way to talk about what is in front of us and the language reinforces itself.
One consequence is the blurring of the description of clinical states and the tendency to overlook discriminations in symptoms that have important effects on treatment and outcome. Treatment is applied to a “specific” process in a manner analogous with an antibiotic to an infection. (That analogy – the specific treatment for a specific affliction – gives the treatment its social aura and potency (Moncrieff, 2008).) Hypersomnia, for example, is different from insomnia and weight loss is different from weight gain. These and many more subtle differences affect risks, treatment and outcome but are not discriminated by the portmanteau diagnosis.
Misunderstanding the nature of diagnosis denies us a clear overview of phenomena and the clinically important associations between them: ‘A main source of our failure to understand is that we do not command a clear view of the use of our words. Our grammar is lacking in this sort of perspicuity [Übersichtlichkeit]. A perspicuous [übersichtliche] representation produces just that understanding that consists in “seeing connections”’ (Wittgenstein, 2001, para 122).
If a diagnosis such as depression is not a specific mental process, there cannot be a treatment that is specific to ‘depression’. There may be treatments that are specific to individual phenomena, but that should not be mistaken for a specific antidepressant medicine. The belief in a specific mental process underlying all depression tempts clinicians to apply treatments without discriminating the different effects of those treatments or discriminating important differences among the problems being treated. When different treatments are applied indiscriminately to an assortment of states that have only family resemblances, the effects will regress to the mean. This may explain the apparent ineffectiveness of many treatments in empirical studies (i.e. no better than placebo and no different from each other) (Fournier et al., 2010) when clinical experience shows them effective.
The urge to see explanation and a unity in the diagnostic language is similar to the problem that Wittgenstein saw for philosophy and he made this injunction:
We must do away with all explanation and description alone must take its place. … These are, of course, not empirical problems; they are solved … by looking at the workings of our language … and to recognise those workings in despite of an urge to misunderstand them. … a battle against the bewitchment of our intelligence by means of our language. (Wittgenstein, 2001, para 109)
Conclusion
A diagnosis is a word in a professional language (or language game) that is understood in the professional language and the activities into which it is woven. It points neither to a fixed phenomenological picture nor to a particular endophenotype or underlying mental process. It varies over time and place according to the needs and predicaments of the profession and patients.
Diagnoses refer to families of cases that are related flexibly and intricately by family resemblance, not by an essence or by a single characteristic or process. They resemble general terms such as ‘games’ that do not have a essential or unifying characteristic. Thus, mental activities such as ‘reading’ or ‘understanding’ and ‘depression’ (in its clinical use) are recognitions of intricate relationships of various human behaviours or experiences with the world. They are not mental processes in themselves.
Psychiatry has taken diagnoses as pointers to mental processes which explain conditions and are the targets for treatment and research. However, it is a misconception to identify many of the cardinal diagnoses in psychiatry with a fixed phenomenological picture or underlying explanatory mental process.
The absence of an underlying mental process gives diagnosis no attachment point for empirical or scientific explanation as it is usually conceived. This makes understandable the repeated failures and non-replication of scientific explanations directed to diagnoses.
This points to the clinical need to replace explanation with description and the elucidation that provides a ‘… perspicuous overview (übersichtlische Darstellung) [that] produces just that understanding which consists in “seeing connections”’ (Wittgenstein, 2001, para 122). It places sophisticated description, not explanatory theorising, at the centre of psychiatric practice.
The question ‘What is depression?’ shares its futility with questions such as ‘What is art?’, or perhaps ‘What is human nature?’ Like many diagnoses it does not point to objects or processes but point to families of human states or actions which have no essence and cannot be explained or confronted by an empirical method.
Footnotes
Acknowledgements
James Durham provided early comment and debate. Joanna Moncrieff (UCL London), Professor Michael Nedo and Dr Matthias Kross (Wittgenstein Archive Cambridge) gave vitally important advice and comment. Rosa Saladino made valuable comment on later drafts.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.