Psychiatry in the Scientific Image is a much anticipated book tackling some of the most fundamental issues of psychiatry using tools from the philosophy of the sciences. Murphy's knowledge of both fields is exemplary; his clear writing and perspicuous argumentation contribute to making this book an ideal introduction (though it is much more than an introduction) to anyone interested in the interface between the aforementioned fields. Though some of the positions defended by the author are hardly new or revolutionary — he is an advocate of the "medical model" of mental illness — the argumentation for them is more rigorous and thorough than most of what you usually find in the literature on the topic. The book being a no-nonsense defense of the medical model, it makes a solid read for that model's friends or foes alike. 

Murphy's book looks at three issues in psychiatry: the definition of mental illnesses (Chapters 2 to 3), their explanation (Chapters 4 to 8) and their classification (Chapters 9 to 10). The book is a reaction to the atheoretical stance towards mental disorders taken since DSM-III[1]. Throughout, Murphy pleads for a reintroduction of science (specifically cognitive neuroscience) in psychiatry — thus the allusion to Sellars' "scientific image" in the title — adding that his "book is deeply reactionary. It is a qualified defense of the medical model, which states that psychiatry is a branch of medicine dedicated to uncovering the neurological basis of disease entities … I think it's especially appealing when one remembers that studying the brain draws on the cognitive and social sciences as well as … molecular biology …." (10). As suggested by the last section of the quote, Murphy is not an average medical model advocate: his model is more ecunimical in spirit, focusing not only on neurological factors as advocates of the traditional model do (Guze, 1992, for example), but also on cognitive and social factors in the explanation of diseases. His version of the medical model is thus "biopsychosocial" (see below).

In Chapter 2 and Chapter 3 (The Concept of Mental Disorder; Psychiatry and Folk Psychology, respectively), Murphy uses Wakefield's (1999) "harmful dysfunction" analyses of mental illnesses as a foil for his own 'objectivist' "two stage picture" of psychiatry. His main claim in these chapters is that conceptual analysis, as practiced by Wakefield, is not the right tool to define mental illness. Like Wakefield, Murphy accepts the idea that the failure of a normal mind to function (dysfunction) is a necessary condition of mental illness, but he thinks that claims about what "normal functioning" is and what parts of the realm of the "mental" are should not rest on folk intuitions, but rather on the 'sciences' of the mental – currently "cognitive neuroscience" (best exemplified by the work of Nancy Andreasen, Christopher Frith and James Blair). Murphy asserts that psychiatry should be viewed as a form of "clinical cognitive neuroscience", a move that produces some counter-intuitive consequences, one of which being the rejection of current psychiatry's demarcation between physical and mental disorders. Murphy rightly points out that part of the reason we consider something a mental disorder rather than a physical one, has to do with historical and political reasons and not with scientific ones. Rather than starting with the set of disorders currently found in nosological manuals (like the DSM) and trying to infer a coherent notion of what the mental is, Murphy thinks it would be better to look at what sciences that are studying the mind are considering as their province and delimiting the domain of disorders from that starting point. The result is the rejection of the divisions between neurology and neuropsychology and psychiatry. "The view of the mind I rely on", Murphy argues, "says that when we talk about the mental we are referring to capacities that enable us to perceive, understand, and act on our environments, among other things. So 'the mental' covers states and processes that play a very direct role in intelligent action, including processes such as perceiving, remembering, inferring, and a wide variety of motivational states" (63). Adopting this view has some counterintuitive consequences as for instance, "[c]ommon sense denies that you have a mental disorder if you are in a diabetic coma, or have no cortical activity because you are on a life-support machine and with only a brain-stem intact. Yet mental functioning is massively impaired in these cases, so I have to call them instances of mental disorders" (62). As we see, psychiatry conceived as a form of clinical cognitive neuroscience has revolutionary potential.

A central element to the two-stage picture of mental disorders is the notion of function and the related notion of "dysfunction". Wakefield proposed that the notion of function used in psychiatry has to be based on evolutionary considerations, an idea that Murphy rejects, as in his view ascriptions of function are not always based on these considerations. Murphy believes that psychiatry should opt rather for what is known as "Cummins-function" where the function of a component is its causal contribution to the working of the whole system, and this, whether or not the component has been selected or designed to execute that function. In that type of model, a mental disorder is produced by the failure of a mechanism (or mechanisms) to perform its (their) function(s) and thus to contribute to the working of the system[2].

I agree with Murphy that intuitive "evolved functions" mentioned by Wakefield are probably not what psychiatry is or should be based upon, but it seems that there is another notion of function that is worth discussing and that is not addressed:  the notion of "propensity function" (note that I am not suggesting that it is a notion that would fair better than "evolved functions"). It is possible that when we are discussing the function and dysfunction of a mechanism that what we are really interested in is the current adaptivity of the behavior outputed by it, or its probable reproductive outcomes. One consequence of propensity function analysis is that, "[b]ecause propensity functions involve the conferring of advantage relative to a specified environment, either actual or potential, categories predicated upon propensity concepts are rather malleable" (Woodfolk, 1999, 664). That notion of function would have the advantage of explaining why the « gourmet syndrome » or « dyslexia » can be considered as « mental disorders » in certain environments or in certain cultures and not in others, so it would fair very well with the social part of Muphy's biosocial model.

Finally, like Wakefield, Murphy thinks that value judgments make the second stage of the picture of mental disorder. Since value judgments can be dependent on the demands of distinct projects, it is possible to end up with different coexisting concepts of mental disorders or that the concepts fluctuate over time or across cultures, depending on these projects. This renders the concept of "… 'mental disorder' [like the concepts of] 'pest', "'weed', or 'vermin'. Weeds and vermin are not 'natural kinds', but they are made up of natural kinds that can be explained empirically" (98). In other words, mental disorders do not form a natural kind, even if they have an objective basis (they are the result of dysfunctions) that can be studied by science.

Chapter 4 (The Medical Model and the Foundations of Psychiatric Explanation) compares two versions of the medical model: a more brain-based model (Kandel's molecular psychiatry) and a more pluralistic, biopsychosocial model (Andreasen's cognitive neuroscientific psychiatry). "These two models agree that mental illness can be caused by an array of biological and cognitive factors. The dispute between their adherents turns on the belief held by proponents of the medical model that molecular neurobiological explanations are more fundamental than other explanations.  … However, although I reject the idea that molecular reductionist theories are more fundamental in general, I do argue that they enjoy a privileged role in some cases. In general, molecular reduction, especially as it involves animal models, gives us a special kind of understanding that is closely tied to our abilities to intervene and manipulate a system in therapy, and to test therapeutic assumptions" (12-13). Adopting a view championed by Schaffner (1993, 1994), models of diseases are conceived as templates to be filled in at different levels of explanation (i.e. genetic, molecular, neurological, cognitive). In some cases, a disorder's features will be explained only by factors at one level of explanation — such as genes —  but it is to be expected that in most cases, the explanation will not be that simple. In these cases, the explanation will typically be crossing levels and will include references to factors at multiple levels of explanation (including social factors, as Murphy argues for in Chapter 7).

I am a bit surprised to find no reference in this chapter to models such as the one proposed by Morton and Frith (1995, 2001).  These psychologists proposed a multi-level model of explanation for developmental disorders exactly of the same sort that Murphy suggests we should adopt, a model that includes environmental causes (both physical and cultural; for example see their explanation of dyslexia). Speaking of developmental disorders, one could have wanted to find a discussion of possible principled reasons why modular models advocated by Murphy might not work for developmental disorders. What Karmiloff-Smith has been arguing for years now is that widespread damage to cognitive functions are not circumscribed ones: "[T]he delation, reduplication or mispositionning of genes [that are believed to be the basis of disorders like autism, Williams syndrome and schizophrenia] will be expected to subtly change the course of developmental pathways, with stronger effects on some outcomes and weaker effects on others. A totally specific disorder will, ex hypothesis, be extremely unlikely, thereby changing the focus of research in pathology …" (1998, 390). Murphy discusses schizophrenia in terms of a "global" disorder of the mind in Chapter 6, but he does not seem to be preoccupied by the effects these kinds of models can have on his "dysfunction analysis" of disorder.

Chapter 5 (The Limits of Mechanical Explanation) presents the limits of the model of explanation advocated in the precedent chapter. The two-stage medical model works for some parts of the mind where what constitutes normal functioning is straightforward, however, it is another story when the capacities of interest have to do with rational behavior. In these cases, malfunction must be identified with reference to a theory of rationality and the norms it prescribes. If one follows Fodor here, if we have a clear idea of how we can mechanize modular (peripheral) capacities like vision or sentence parsing, we do not have the faintest idea of how we could mechanize central capacities like analogical reasoning or abduction. In other words, for central capacities, we cannot establish a judgment of malfunction just by looking at the system; for that we need a normative theory. The brain of someone suffering from delusion might look different from my brain, but it is not sufficient to conclude that their brain is malfunctioning — for this we need a theory of what a 'normal' person would infer when having a certain deviant experience, such as hearing a voice in their head. According to Murphy, one can naturalize reason (that is, one can provide a story concerning the possible evolved epistemic function of rationality), but not mechanize it. The consequence of this is that "[t]he assessment and treatment of these disorders [like delusion or alcoholism] usually requires a inexplicit model of what good reasoning is. If rationality is a normative notion, the model must be in part normative. Since this model is normative, it undermines the pretensions of the medical model to purely positive science" (153). However Murphy reminds us that this concession is not an endorsement of the anti-psychiatry's position of people like Szasz, because it applies only to a limited number of mental illnesses, and the norms in question are not social norms, but rather rationality norms.

Chapter 6 (A More or Less Realist Theory of Validation as Causal Explanation) tackles the problem of validation. According to Murphy, scientific psychiatry should "[…] find, among the variety, an idealized representation of the symptoms and course of a mental illness, and try to uncover the idealized causal relationships that account for that idealized picture. I call the idealized representation of the syndrome and its course and "exemplar" of the disorder, and causal validation is the explanation of why exemplars have the form they do. The result is a model, consisting of the exemplar plus its causal backstory" (202). This idealized version of the disease and its causes, once built, can be used in clinical psychiatry. In clinical psychiatry, one compares actual cases with idealized types and looks for a fit. This realist's view advocated by Murphy contrasts with the DSM's way of validating its constructs. The DSM has an operationalist's view of validation; in order for a construct to be validated, symptoms should be found to reliably vary with diseases. As Murphy puts it, "The point of construct validity is not to uncover something in nature, but to find agreement across measurements" (219). Murphy recognizes that operationalist validation is not pointless, as long as it is not conceived as the end of psychiatry. Following Hempel (1965), Murphy thinks that construct validity can indeed be used heuristically to discover "real" mental illness.

In the next two chapters, Murphy considers two possible extensions to his modified version of the medical model. In Chapter 7 (Social Construction and Sociological Causation), he militates for the integration of social and cultural factors in the explanation of mental illnesses. Chapter 8 (Evolutionary Explanations of Psychopathology) draws a darker picture of the prospects of evolutionary considerations in the explanation of mental disorders – but let's start with Chapter 7.

Social constructionists are divided into two camps: the skeptics and the realists. Skeptics (Foucault and Szasz, for example) are usually interested in highlighting that what we call mental illnesses are in fact deviant behaviors shunned by society. Realists are usually anthropologists who point to the role of social and cultural factors in the constitution of mental illness. If the skeptic position encounters numerous problems (relativism, for one) and cannot really be integrated to the medical model; the realist position, on the other hand, fair much better  according to Murphy. As Murphy shows, cases like anorexia, dysthemia, latah (Murphy treats the latter as a mental illness, though it is not clear that those suffering from it think of themselves as or are treated like the mentally ill, see Barholomews 2000) all necessitate a reference to social and cultural factors in their etiologies. This should not be problematic for a realist's model of psychopathology explanation as long as there is a mechanism that can explain how these factors can influence or cause pathologies — so as long as we can produce a causal link of a similar type to instances of genes or lesions. Murphy claims that social factors can indeed be mediated by mental representations; he writes, "For social explanations to play a genuine explanatory role they must work via a material mechanism. This can explain the content of particular symptoms (the particular form that delusions take is [an] example) and the epidemiological patterns exhibited by mental illnesses. Social construction, understood in my broad sense, is not limited to the explanation of unstable phenomena, nor is it anti-realist or incompatible with natural explanations. Both social and natural explanations can be integrated in the same neuropsychological structure, which suggests that debates about whether mental disorders are natural or social kinds are beside the point: they can be natural kinds even if part of their explanations are social." (279-80)

Murphy sometimes writes as if the only way through which we could address social construction are instances of when social or cultural factors have an effect on a psychopathology through mental representations. For instance, he writes: "The obvious place to turn for a mechanism of social causation is mental representation" (256). I personally think we could also speak of social construction in cases where social factors act in a more indirect way on pathology. For instance, Panksepp (1999) have proposed that one cause of ADHD might be the fact that in our society, children cannot play as much as they need too. ADHD would then be caused by our ideas about how children should behave and what they should do during their day (for example: sitting quietly in school). This explanation of ADHD might be false, but it illustrates the possibility of a process that we could brand "social or cultural niche construction" and that could also qualify as social construction, even though it does not necessarily implicate any mental representations from the disordered patient in the causal process. Finally, there is one conclusion from his position that Murphy does not consider: if what he says is right, it should have major effects on nosologies. One of these effects, upon which psychiatric anthropologists sometimes insist, is that this conception of psychopathology might force us to adopt local nosologies and reject transcultural ones.

If Murphy and Stich (2000) had convinced you of the fact that we were at the dawn of evolutionary psychiatry, you might find yourself wanting to take a pause after reading Chapter 8. Looking at paradigmatic examples of evolutionary psychiatry –psychopathy, depression, schizophrenia and anxiety disorders– Murphy concludes that: "In general, the evolutionary explanations … fail to stand up. The problem typically is that they fit only some symptoms of a disorder — psychopaths seem too irrational to be successful long-term manipulators, depressed people seem too unmotivated to be plotting a new strategy or too miserable to be reconciled with their lot, and schizotypal patients seem short of leadership skills. The failure is one of form not matching alleged function. In the case of some types of anxiety, though, the match is there, and these may be plausible candidates for a direct mismatch explanation, especially if fleshed out with assumptions about the distribution of anxiety in a population" (304-5).

The logic of Chapter 8 is far from clear to me. Murphy is right, evolutionary psychiatry is presently in a poor state; it might just be the impression I got after reading the chapter, but it feels as if he is more interested in criticizing evolutionary psychiatry than in proposing ways in which it could constitute a meaningful addition to his model. The adoption of an evolutionary attitude could for instance serve a heuristic role in identifying functions that the mind has had considering the kind of problems it had to solve in its adaptive environment (but is this compatible with his stance on function adopted in chapter 3. If not, the role of evolutionary psychiatry in Murphy's cognitive neuroscientific psychiatry seems hard to understand). After all, evolutionary psychologists strive to go beyond intuitions and instinct blindness (the manifest image that Murphy claims we should distance ourselves from) in order to produce a scientific picture of the normal mind's organization.

Chapters 9 and 10 (Classification; Classification in Psychiatry, respectively) were slightly less interesting to me. Once you understand the model of explanation proposed by Murphy in the mid-section of the book, it seems that the solution to the classification problem in psychiatry is obvious: construct your classification so it reflects (or allows you to reflect) the causal knowledge you have (or will have) gathered through the sciences that are devoted to studying the mind. This contrasts with what the DSM is explicitly claiming to do. As Spitzer noted: "The authors of DSM-III sought to achieve this agreement by separating psychiatric observation from psychiatric theory. The common classification scheme would consist of categories whose meanings could be defined as far as possible through direct observation. In this way the adherents of different schools could nonetheless agree on basic terminology because disputes regarding definitions could be settled by appeal to what all could observe and could not reasonably deny …" (1995, 92). Murphy's solution to the problem of relying on « clinical phenomenology » or « descriptive psychopathology » is to turn to his ecumenical version of cognitive neuroscience for an understanding of how the normal mind works and how it can malfunction. The incomplete knowledge provided by these sciences should not be an obstacle to the adoption of a scientific framework. As Murphy remarks, one can distinguish between two situations « … causal discrimination and causal understanding. Causal discrimination is possible when we have a syndrome that has a cause we know to be distinct from the cause of a different syndrome even if we are unsure about the nature of the causes in each case. Or, we might have good explanations of the two causal relationships; that is, causal understanding. The claim I am making is that classification can reflect causal discrimination in the absence of causal understanding. Causal discrimination, as I see it, occurs when enough information exists to generate a reasonable set of expectations about causal topography, even if the details are still hidden. We might be able to classify disorders in gross classes even without clear details of the causal story in each case" (319). Causal discrimination — which can be achieved through lesion studies, pharmacological dissection or psychological tests, even if none of these methods is a hundred percent full proof — is more than enough to build a solid nosology.

 In conclusion, Murphy's book is a very good example of what cutting-edge philosophy of (special) sciences should look like. Though I think that the general argumentative line of the book would have been stronger without Chapters 5 and 8, which are more focused on the limits of the medical model than on presenting and defending the model as such, the book is a must-read for anyone interested in either or both of the fields of philosophy of science or psychiatry.

Many thanks to Pierre Poirier and Amanda Cox for their comments on this review.

Bibliography

Bartholomew, R. 2000. Exotic Deviance: Medicalizing Cultural Idioms From Strangeness to Illness, Colorado University Press.

Guze, S. B. 1992. Why Psychiatry is a Branch of Medicine,Oxford, Oxford University Press.

Hempel, C. 1965. "Fundamentals of Taxonomy". In Aspects of Scientific Explanation, NY: Free Press, 137-154.

Karmiloff-Smith, A. 1998. "Development itself is the Key to Understanding Developmental Disorders". Trends in Cognitive Sciences,2, 10, 389-398.

Kirk, S. A. and H. Kutchins. 1992. The Selling of the DSM: The Rhetoric of Science in Psychiatry. New York: Aldine Transaction.

Morton, J. and U. Frith. 1995. "Causal Modeling: A Structural Approach to Developmental Psychopathology". In: D. Cicchetti and D.J. Cohen, Editors, Developmental psychopathology, Vol. 1, John Wiley & Sons, New York, 357–390.

— 2001. "Why We Need Cognition: Cause and Developmental Disorder". In : E. Dupoux (ed.), Language, Brain and Cognitive Development, Cambridge : MIT Press,  263-278.

Murphy, D. et S. Stich. 2000. "Darwin in the Madhouse: Evolutionary Psychology and The Classification of Mental Disorders". In P.Carruthers and A.Chamberlain (eds.),  Evolution and the Human Mind, Cambridge, Cambridge University Press, 62-92.

Panksepp, J. "Attention Deficit Hyperactivity Disorders, Psychostimulants, and Intolerance of Childhood Playfullness: a Tragedy in Making". Current Directions in Psychological Science,1999, 91-98.

 Prinz, 2006 "Is the Mind Really Modular?". In R. Stainton (Ed.), Contemporary Debates in Cognitive Science, Oxford: Blackwell, 22-36.

Schaffner 1993. Discovery and Explanation in Biology and Medicine. Chicago: Chicago University Press.

1994 . "Psychiatry and Molecular Biology: Reductionistic Approaches to Schizophrenia". In : J.Z. Sadler, O. P. Wiggins et M. A. Schwartz (eds), Philosophical Perspectives on Psychiatric Diagnostic Classification,Washington, John Hopkins University Press, 279-294.

Shallice, T. 1988. From Neuropsychology to Mental Structure. Cambridge: Cambridge University Press.

Wakefield, J. C. 1992. "Disorder as Harmful Dysfunction: a Conceptual Critique of DSM-III-R's Definition of Mental Disorder". Psychological Review, 99, 232- 247.

 Woolfolk, R. 1999. "Malfunction and Mental Illness". The Monist, 82, 4, 658-670.

 

---