No, Psychiatric Diagnoses Do Not Reflect “Circular Logic”


Not knowing the pathophysiology of a condition does not mean we have no causal explanation of the patient’s suffering and incapacity.


Olivier Le Moal/AdobeStock


Imagine that you and your traveling companion are touring the ruins of the ancient city of Pompeii. Your historically uninformed companion asks you, “So, what caused the destruction of Pompeii?” You reply that an erupting volcano in 79 AD was the cause, and go on to describe its features: the expulsion of gases, rock fragments, and molten lava spewing from within the Earth through a vent onto the Earth’s surface. Your companion strenuously objects, “No, no! All you did was describe the observable features of a volcano. I want to know the cause of Pompeii’s destruction!” At that point, you might be puzzled or a bit annoyed, and reply that you have just identified the cause.

But maybe your friend was trying to pose a deeper question, such as, “What is the cause of volcanoes?” You see the problem: you were answering by invoking 1 level of causal explanation (level 1), whereas your friend was probably seeking a level 2 causal explanation. The latter is a perfectly legitimate and laudable mode of inquiry—but it remains the case that the destruction of Pompeii was indeed caused by an erupting volcano. There is nothing wrong with the level 1 explanation, incomplete though it is. Often, in science and medicine, that is the best we can do in our present state of knowledge.

Now, transposing the argument into the realm of psychiatric diagnosis: we often hear critics argue that psychiatric diagnoses are “purely descriptive” and are “currently defined only by symptoms. That is, they do not refer to any known pathophysiological processes or specific causes.”1

These critics argue that DSM diagnoses are merely “agreed-upon labels—a kind of shorthand—for describing symptoms…”2 This claim often leads to the charge that psychiatric diagnoses are little more than exercises in circular logic. Why? Because, the argument goes, “…a purely descriptive diagnosis cannot be the cause for its symptoms, because it merely describes them: [for example], depression cannot be the cause of depressed mood.”1

As one prominent critic put it2:

“Here is the circular logic: How do we know a patient has depression? Because they have certain symptoms. Why are they having these symptoms? Because they have depression.”

The argument then expands to allege that “misleading circular causal claims” are potentially harmful, because they may lead the public “to misunderstand the nature of mental health problems.”1

To be clear: we acknowledge many problems and inadequacies in current DSM-5 psychiatric diagnostic categories, including but not limited to the DSM’s rejection of diagnostic hierarchies and the problematic heterogeneity of the “mental disorder” construct.3 Nevertheless, we believe the critics’ circularity argument is ill-founded, misleading, and fallacious.

To put it simply, these critics are insisting on a level 2 causal explanation while wholly discounting the (admittedly limited) value of a level 1 causal explanation. We will show presently why diagnoses like schizophrenia or bipolar disorder, properly made, are neither circular nor uninformative nor lacking in explanatory value—despite the absence of “known pathophysiological processes” as part of their diagnostic criteria. The critical phrase here is “properly made.” First though, we need to clear away some rhetorical underbrush that often obscures the reality of psychiatric diagnosis.

Misunderstanding the DSM-5

In our experience, very few US psychiatrists and psychotherapists have taken the time to read through the first 25 pages of the DSM-5; ie, the “Introduction” and “Use of the Manual” sections. This has led to several mistaken beliefs about the nature, scope, and purpose of the diagnostic categories—and, more important, of the diagnostic process.

First off, it is simply not the case that the DSM diagnostic categories are defined only by symptoms, understood as the patient’s subjective report of what he or she feels or experiences.4 On the contrary, many psychiatric diagnostic categories include a variety of signs as part of their criteria set; ie, objective features that can be clinically observed and measured.5 Examples include significant weight loss (in major depressive episode, anorexia nervosa); psychomotor agitation or retardation; stereotyped movements; distractibility (during evaluation), avoidance of eye-to-eye gaze (in autism); and, of course, numerous objectively measurable deficits in cognitive function, such as impaired recent memory, inability to calculate, etc. We would also classify observable features like pressured speech, loose associations, and markedly elevated or depressed affect as signs—not symptoms.

Secondly, in order to make a valid DSM-5 diagnosis, the clinician must—yes, must—devise a case formulation based on biopsychosocial factors6:

"…that may have contributed to developing a given mental disorder. Hence it is not sufficient to simply check off the symptoms in the diagnostic criteria to make a mental disorder diagnosis."

In our experience, this DSM-5 requirement is rarely appreciated, much less satisfied, by most US clinicians. This is unfortunate, since the case formulation very clearly provides substantive, explanatory content for the patient's condition—not merely a recitation of the patient’s symptoms.7

Finally, the diagnostic process of the DSM-5—apart from its diagnostic criteria—requires a series of rule outs before a final diagnosis can be made. For example, a diagnosis of schizophrenia (Criterion E) requires that “…the disturbance is not attributable to the physiological effects of a substance…or another medical condition.”6 In clinical practice, this requires the psychiatrist to consider and rule out a variety of neurological, endocrine, and infectious conditions, ranging from complex partial seizures to “myxedema madness” (severe hypothyroidism) to tertiary syphilis—causes of so-called “secondary psychosis.”8

The Circularity Fallacy

Critics who claim that psychiatric diagnoses are merely tautologies do not seem to realize that what the patient does not have is just as important—and as informative—as what they do have. As our colleague, Awais Aftab, MD, has pointed out9:

“An important thing about diagnoses in medicine is that they don’t just identify something, they also typically exclude other things. For example, to say “you have unipolar depression”, also implies, “I don’t think you have bipolar depression”. To say, “you have generalized anxiety disorder” also means “I don’t think your anxiety is secondary to psychotic symptoms”. In this sense, to say “Your anxiety is caused by generalized anxiety disorder” can be interpreted as “The way I understand your anxiety, it seems to be best described as 'generalized anxiety disorder' rather than 'major depressive disorder with anxious features', or as 'panic disorder'… We sometimes colloquially use ‘caused by’ with the meaning of ‘this is how it best makes sense’ or ‘this is how it is best described.’”

Thus, when we provide the patient with a psychiatric diagnosis, we are simply hypothesizing the existence of a condition that “best makes sense” of the patient’s presenting signs and symptoms. This is very close to what philosophers of science call “inference to the best explanation.”10 In so doing, we are not making any metaphysical claims, or reifying the condition by positing some essence, substance, or thing residing inside the patient—akin to, say, a burst appendix.

Indeed, we recognize that psychiatric (and medical) diagnoses are in large measure socially constructed—but that does not render them in any sense unreal, metaphorical, or tautological. (The concept of health is also socially constructed, but few would argue that health is thereby rendered mythical, metaphorical, or in some sense nonexistent).

That said, Dr Aftab points out that there are some diagnoses—perhaps we should call them pseudo-diagnoses—that embody so little descriptive or causal content as to be trivially non-explanatory. He gives the example of “fever of unknown origin” (FUO). As Aftab puts it9:

“Imagine saying to a patient, ‘Your fever is caused by FUO.’ That makes no sense at all. We know that this specific fever has some unidentified cause, yet that cause is so abstract, so distant, steeped in so much ignorance, that the mere knowledge that there has to be some cause doesn’t make it so that the mere use of the term offers us some causal explanation.”

In contrast, we would argue that the conditions often termed serious mental illnesses do not fit the FUO paradigm.11 On the contrary, debilitating afflictions like schizophrenia, bipolar disorder, and melancholic/psychotic forms of major depression have so many external validators—eg, family history, course of illness, characteristic biomarkers, genetic risk factors, and typical response to treatment—that it makes sense to speak of these conditions as causing symptoms. Indeed, no less a source than the Substance Abuse and Mental Health Services Administration (SAMHSA) notes that11:

“Bipolar Disorder is a brain disorder that causes intense shifts in mood, energy, and activity levels…[and] schizophrenia is a chronic and severe mental disorder that causes people to interpret reality abnormally.” (Italics added).

There is nothing circular or tautological in SAMHSA’s statements. Essentially, SAMHSA is invoking what we have called level 1 causality, cognizant that we do not know the precise pathophysiological processes that explain bipolar disorder or schizophrenia (level 2 causality). It is just as reasonable to state that “Smith’s command auditory hallucinations and paranoid delusions are caused by his having schizophrenia” as it is to state that “The destruction of Pompeii was caused by an erupting volcano.” The truth of the first claim is not dependent on knowing the etiopathology of schizophrenia.12 The truth of the second claim is not dependent on any knowledge of, say, shifting tectonic plates in the genesis of volcanoes.

Ironically, the paper by Kajanoja and Valtonen demonstrates how the circularity charge collapses under its own weight. They write (footnote 3) that1:

"…it is logically circular to say that depression caused the person’s depressed mood, [but] it is not circular to say that depression caused the person to not attend an event." (italics added).

Indeed, we completely agree. But this is precisely where the circularity charge falls apart. For if the depression itself can cause a person to "not attend an event," then, ipso facto, depression itself necessarily has causal efficacy; ie, it can itself act on individuals to produce clinical or behavioral effects. Consequently, there is no reason in principle why depression could not also cause the person to sleep poorly, eat poorly, feel guilty or suicidal, or experience psychomotor slowing. And importantly: none of these facts requires us to have a complete or even a partial understanding of the pathophysiology or etiopathology of depression itself.

The Historical Identification of Disease

Critics alleging the circularity of psychiatric diagnosis seem to have limited understanding of the identification of disease states through the ages. Indeed, the history of medicine is replete with examples of diseases first identified long before their underlying physiological mechanisms—their etiopathology—were known. In 1817, when James Parkinson described a well-defined constellation of signs and symptoms he called “the shaking palsy,” he had in fact identified a disease state that caused the patient’s signs and symptoms—and what would later become known as Parkinson disease. Importantly, he did so without knowing anything at all about what caused the disease that would bear his name. It would be absurd, on that basis, to say that Parkinson disease did not exist—or was not the cause of parkinsonian symptoms—until its etiopathology was discovered in the 1950s and 60s. In fact, the French neurologist Jean-Martin Charcot (1825–1893) explicitly referred to the condition as Parkinson disease as early as 1888—decades before its etiopathology was uncovered.13 This fact flatly contradicts the position that only conditions with known and identified pathophysiology count as bona fide diseases.

Finally, to be clear: in asserting that some psychiatric diagnoses can legitimately be said to be causative of symptoms, we are not claiming that psychiatric disorders arise solely from constitutional-biological factors. On the contrary: we recognize that powerful psychological and social forces often set in motion a pathological process that results in the manifest symptoms. This fact reaffirms the importance of the case formulation and its biopsychosocial orientation.

Concluding Thoughts

Causality is not a simple, single-layered concept. Aristotle, for example, recognized 4 types or levels of causation, which he called material, formal, efficient, and final.14 Similarly, in our response to Dr Aftab,15 we described 3 alternate uses or senses of the term causality:

  • Etiopathological causation: this refers, ultimately, to the biological and physico-chemical mechanisms through which a disease process develops
  • Clinical causation: this refers to the clinician’s identification of a good fit between the patient’s presenting signs and symptoms, and a recognized clinical entity
  • Causation in ordinary language: this refers to the way the average person often uses the term “cause,” or its derivative, “because” (by cause).For example, “The doctor said that Joe gets palpitations, dizziness, and sweating because Joe has panic disorder.” And as the philosopher Ludwig Wittgenstein observed, “Ordinary language is all right.”16

To be sure, our current psychiatric nosology leaves much to be desired. One of us (RP) has quipped that the DSM-5 is “by far the worst diagnostic framework psychiatry could have chosen—except for all the rest.”7 Of course, we all wish that psychiatry were at the point of invoking level 2 causality in our diagnoses, and research continues to yield important knowledge in the realm of pathophysiology.17 But our incomplete understanding does not mean that our diagnostic categories are merely short-hand labels for a bunch of symptoms—or empty, acausal tautologies. When arrived at carefully—including a biopsychosocial case formulation—a psychiatric diagnosis can explain a great deal about the patient’s presenting problem, and remains the royal road to effective treatment. 

Acknowledgment: Dr Pies would like to thank Jani Kajanoja MD, PhD, for his stimulating and collegial email exchanges related to this topic. We also thank Awais Aftab, MD, for hosting some of our exchanges.

Dr Pies is Professor Emeritus of Psychiatry and Lecturer on Bioethics and Humanities, SUNY Upstate Medical University; Clinical Professor of Psychiatry Emeritus, Tufts University School of Medicine; and Editor in Chief Emeritus of Psychiatric Times (2007-2010). Dr Pies is the author of several books. A collection of his works can be found on Amazon. Dr Ruffalo is an assistant professor of psychiatry at the University of Central Florida College of Medicine in Orlando and adjunct instructor of psychiatry at Tufts University School of Medicine in Boston, Massachusetts.


1. Kajanoja J, Valtonen J. A descriptive diagnosis or a causal explanation? Accuracy of depictions of depression on authoritative health organization websites. Psychopathology. 2024:1-10.

2. Shedler J. A psychiatric diagnosis is not a disease. Psychology Today. Updated May 17, 2024. Accessed June 21, 2024.

3. Ghaemi SN. After the failure of DSM: clinical research on psychiatric diagnosis. World Psychiatry. 2018;17(3):301-302.

4. Symptom. National Cancer Institute. Accessed June 21, 2024.

5. Pies RW. Misunderstanding psychiatry at the highest level. Psychiatric Times. September 9, 2011.

6. Diagnostic and Statistical Manual of Mental Disorders. 5th edition. American Psychiatric Association; 2013.

7. Pies RW. Poor DSM-5—so misunderstood! Psychiatric Times. March 23, 2021.

8. McKee J, Brahm N. Medical mimics: differential diagnostic considerations for psychiatric symptoms. Ment Health Clin. 2016;6(6):289-296.

9. Aftab A. Can symptoms be caused by descriptive syndromes? An analysis. Psychiatry at the Margins. Accessed June 21, 2024. 

10. Inference to the best explanation. The Information Philosopher. Accessed June 21, 2024.

11. Living well with serious mental illness. SAMHSA.

12. Ruffalo ML, Pies RW. What is meant by a psychiatric diagnosis? Psychology Today. June 7, 2020. Accessed June 21, 2024.

13. Goetz CG. The history of Parkinson's disease: early clinical descriptions and neurological therapies. Cold Spring Harb Perspect Med. 2011;1(1):a008862.

14. Falcon A. Aristotle on causality. In: Zalta EN, Nodelman U, eds. The Stanford Encyclopedia of Philosophy (Spring 2023 Edition).

15. Aftab A. An exchange with Ruffalo and Pies. More on diagnosis and diagnostic explanation. Psychiatry at the Margins. June 12, 2020. Accessed June 21, 2024.

16. Wittgenstein L. The Blue and Brown Books. Harper & Row; 1958.

17. Luvsannyam E, Jain MS, Pormento MKL, et al. Neurobiology of schizophrenia: a comprehensive review. Cureus. 2022;14(4):e23959.

Recent Videos
Dune Part 2
© 2024 MJH Life Sciences

All rights reserved.