Abstract
Radical objectivity is only intelligible and accessible through radical subjectivity. This paradox has been much commented on by Heidegger… and by Merleau-Ponty.
—Charles Taylor The moral topography of the self
Radical uncertainty remains inherent in medicine's status as a science of individuals.
—K. M. Hunter Doctors’ stories: the narrative structure of medical knowledge
A formative experience in the training of doctors, including psychiatrists, is the transition from the lecture hall and laboratory class to the patient's bedside. This has been described as a descent from an ordered world of precision and rationality into a maelstrom of uncertainty and ambiguity, and has long been recognized as a major stressor in the doctor's personal and professional development [1], [2]. One reason, and indeed a central one, why this experience is so unsettling lies in a widespread misunderstanding of the nature of science and its methods, particularly the role of observation and data-gathering in producing scientific knowledge as it is presented in the medical curriculum, including postgraduate psychiatric training. Let us begin with an overview of the role of observation in science.
The history of the sciences that are foundational to psychiatric education demonstrate that a struggle with ambiguous observations and competing, equally plausible interpretations of phenomena, arrived at by different methodologies, is part of the reality of scientific practice. Appreciating this may serve as a corrective to the unquestioning acceptance of claims made in the name of ‘evidence'in clinical psychiatry, and may better prepare psychiatrists for the uncertainties of clinical work.
In his influential historiography of the physical sciences, Thomas Kuhn [3] noted how little of the struggle with anomalies and rival theories is taught to students of a scientific discipline, who are therefore ignorant of the tradition of debate among various viewpoints in their field. Instead, students are taught that science provides a single and true account of the world based on observation, which enables formulation of causal laws linking the observations in a logical, mathematically expressed order. These observations are held to be neutral as regards any particular theory or set of values. It is the doctor's burden to leave such certainty behind when treating a patient; the art of medicine rather than its science is the means to resolve the inconsistencies and unpredictability of clinical reality [4].
These popular views are derived from the principal tenet of empiricism, which claims that rational beliefs are either observational or the conclusions of sound argument inferred from observational premises [5]. This account of science is widely held in academic and broader circles and considers observation to be unproblematic or selfevident [4]. In psychiatry it provides the intellectual framework for the DSM classification, of which the implicit cognitive and evaluative foundations remain unexamined [6], [7]. This DSM, in turn, is the foundation on which evidence-based claims are made in psychiatry.
Observation is theory-laden
The impossibility of making theory-free observation in science has long been understood. In the late 19th century, the French mathematician and physicist Pierre Duhem argued that although science enables large numbers of empirical facts to be related to one another according to a comparatively small number of principles, the observations themselves cannot be carried out by a naïve observer; only someone trained in the particular field who has understood its theories, concepts and methods can make the observations [8]. Furthermore, any given observation was the result of a large number of largely unarticulated theories, hypotheses, suppositions and inferences that guided the method of selecting and interpreting the observation. Duhem argued that physicists can never subject an isolated hypothesis to experimental test, but only a whole group of hypotheses; when the experiment is in disagreement with their predictions, the experiment does not designate which hypothesis within the group is faulty [8].
This point was elaborated by Quine [9], who argued that a theory in science (or theory statements) may be thought of as a ‘web of belief'or a ‘field of force’, which interacts with experience at its boundaries. The theory to be tested is inextricably linked with many other theories, including those that subtend the workings of the instruments used to make observations, assumptions and axioms about what constitute relevant observations and the observation language. Any observation that appears to disprove a theory does not indicate what (theoretical) nodal points or connections in the ‘web'are in error. This remains so even if one tries to obey Popper's methodological rule of subjecting rival theories to maximal possible falsification [10].
A further development in this pluralistic view of the relation of theory to observation was proposed by Nelson Goodman, who argued that ‘the world'could be construed in a variety of ways depending on our particular purposes and interests [11]. Goodman argued that there are, in effect, as many ‘worlds'as there are viable constructions of ‘it’(‘it'or ‘the world'entailing no specific reference). Each particular construction proceeds according to its own rules and is interpreted according to the system that determines the rules of construction. Thus, for example, a new metaphor is not just a transposition of words, but may change the way we perceive the world.
These ideas were integrated into the concept of the ‘paradigm'proposed by Thomas Kuhn in his historical account of theory change in the actual practice of science [3]. A paradigm describes ‘the conceptual box into which scientists try to fit nature’: the professional values, methods, theories and concepts that define the work and disciplinary commitments of a group of scientists. Within a particular paradigm, scientists practice what Kuhn termed ‘normal science’, which explores and articulates the paradigm according to the methods prescribed by the paradigm. In so doing, both its usefulness and limitations become evident.
As dissatisfaction with the paradigm grows because of its inadequacies and anomalies, the field enters a stage of ‘crisis'in which the scientists debate the foundations and axioms of their discipline. At some point a revolution occurs and a new paradigm emerges that solves some of the major anomalies in the old paradigm and rearranges the field in a ‘gestalt switch’, which introduces new problems, new methods, and often a new political and pedagogical order. The new paradigm allows the discipline to resume its practice of ‘normal science'and the old paradigm is forgotten.
Duhem, Quine, Goodman and Kuhn are but a few of those who have challenged the traditional views of empiricism and argue against the possibility of simple, ‘objective'observations in science. The neuroscientist and philosopher Churchland summarized the argument with the contention that the empirical adequacy of any theory is both radically under-determined by any evidence that might be available (i.e. the choice of a theory cannot be justified by appeal to empirical observation alone), and also requires that ampliative inferences be made, that is, there are a plurality of theoretical moves that have to be made beyond the observation statements themselves to explain any particular set of observations [12].
According to Churchland, observation is radically contingent upon temporal, neurophysiological and social factors that regulate scientific activity and cannot provide a secure foundation for the tenets of empiricism. Churchland considers science to be a human activity concerned with the acquisition not only of facts about the world, but also of skills that enable a particular formalism to be learned and applied, including applications of this formalism to the very methods and instruments by which these skills are acquired. This view of the selfreflexive capacity of science is similar to the conclusions reached by a very different tradition, namely Husserl's phenomenology.
Phenomenology
‘Phenomena'w as the term used in 18th-century European philosophy to describe the way things are perceived by the (human) senses. Phenomenology, as initiated by the mathematician and philosopher Edmund Husserl was concerned with the nature of appearances and how the distinction may be drawn between appearances and reality [13]. Science, according to Husserl, glossed over this distinction by the use of mathematics, which gave an idealized, abstract picture of the world, what he termed ‘the mathematization of nature’. Useful as this often was, Husserl feared it also took people away from the basic ground of experience, including sensory experience, leading to idealized (and false) distinctions between mind and body, theory and practice, or moral and pragmatic reasoning.
Husserl adopted Brentano's principle of intentionality, that is, that there is no such thing as consciousness, there is only consciousness of something; consciousness is an activity that is directed by intention. This intentionality reflects both the biological endowment of individual human beings and their membership of a culture that shapes perception. This allows for a plurality of lifeworlds for different human beings in different cultures at different historical periods, although Husserl also believed that an exacting study of the suppositions upon which all perceptions were based would lead to a set of common characteristics, which would enable both the natural and human worlds to be studied by a science that would understand its own activity from a human point of view.
Writing his last works under the growing shadow of fascism in Europe, Husserl mourned the loss of the Enlightenment values of rational and free inquiry and their replacement by a bureaucratized system of mathematicized, dehumanized technology. Turning now from the philosophy of science to its history further reveals the limitations of empiricism and the distorted view of science it promotes.
Physical sciences
A well-known story in the Western scientific tradition tells of Galileo, who signalled the birth of the scientific age by dropping two weights off the tower of Pisa and observed that contrary to the time-honoured but hithertountested doctrine of Aristotle, the heavier weight did not land before the lighter one. Introductory physics texts credit Galileo's experiment with sweeping away centuries of Aristotelian sophistry and substituting the authority of rational conclusion derived from observation. It is therefore disconcerting to read the views of contemporary historians of science that Galileo is unlikely to have performed the Pisa experiments in this way and probably derived his laws by a chain of logical arguments that relied on several of Aristotle's ideas [14], [15]. Radical as Galileo's concepts were, they did not arise from simple observational or experimental challenges to an unwieldy, faltering theory, as is popularly described.
Awaiting the results of the 1914 eclipse expedition, which was to test his theory of general relativity, Einstein wrote: ‘I do not doubt anymore the correctness of the whole system, may the observations of the eclipse succeed or not. The sense of the thing (Vernuft der Sache) is too evident’[16]. Einstein's intellectual biographer notes that for Einstein, in addition to its logical and empirical dimensions, a scientific explanation was guided by ‘themata’, which are pretheoretical suppositions about nature, for example, that natural forms are symmetrical [17]. Thus it was the asymmetry of Maxwell's field equations that prompted Einstein's unease, not, as is often claimed, the puzzling conclusions about the nature of light implied by the Michelson– Morley experiment. In his correspondence with the mathematician Ernst Mach, a champion of empiricism, Einstein expressed his growing disenchantment with that doctrine and a corresponding interest in aesthetic factors governing the construction of a theory: simplicity, integrity and respect for the subject [16].
This is similar to what the philosopher of science Polanyi called ‘tacit knowledge’[18]. This is a form of knowledge that resides in what he called ‘subsidiary awareness'and arises from deep immersion in a subject, leading to a non-verbal ‘sense'for the subject.
Churchland believes that the choice of a particular theory on account of its simplicity, coherence and explanatory power is not merely an aesthetic judgement, but reflects an evolutionary process by which the human brain has developed the means of distinguishing relevant information from background ‘noise’[19].
Many more examples could be selected from the physical sciences, but the point is to place a caveat on the aspirations to and the sanctity of making ‘pure'observ ations in science, regardless of whether they are used as ‘evidence'to verify or falsify a theory. The history of science demonstrates that observation is problematical and not the simple, unbiased registering of sense-data.
Biological sciences
Turning to the biological sciences and leaving aside specific conceptual and methodological problems in the scientific study of living organisms, criticisms of empiricism also apply. For example, to claim that a theory of evolution can be derived from ‘pure'observ ational evidence neglects the controversy of many years between followers of Darwin and those of Linnaeus. The former group claimed that the fundamental unit of evolution was the individual and species did not really exist in nature except in the minds of the classifiers, whereas the latter maintained ‘natural populations'di vided into clearly delineated species to be the focus of the evolutionary process [20]. Such differences reflect conceptual disagreements that cannot by settled by mere observation, and indeed determine what observations, utilizing what methods, are to be considered relevant.
When biological science addresses the changeable complexities of its subject matter, at least four aspects of ordering of phenomena are discernible, each with its temporal relations, theories, constructs, models and investigatory techniques. As a pioneer of modern biology, J. B. S. Haldane argued no single level is sufficient in itself to explain biological phenomena [21]. Instead, the four aspects, physicochemical (biochemical), physiological (functional), developmental (morphogenetic) and evolutionary (adaptational), must be integrated by the exercise of sophisticated thinking and disciplined imagination, which far exceeds the empiricist's position.
Furthermore, Haldane's adaptational level admits ethical considerations into biological science as reflected in the eugenic debates around the turn of the 20th century [22] and by the current controversies over sociobiology and its offshoot, evolutionary psychology [23]. The debate cannot be settled by simply appealing to observation made from a supposedly neutral, atheoretical position.
Clinical sciences
Pathology
The theories of disease current in the 19th century serve to illustrate the inadequacies of empiricism in accounting for the development of knowledge. Carl Rokitansky, a leading Viennese pathologist in the 1830s, sought to differentiate between the functions of the pathologist and the clinician, each of whom would make observations in the appropriate setting (mortuary or ward) and jointly contribute to the understanding of disease [24]. Rokitansky's department performed hundreds of autopsies annually, and on the basis of nakedeye and microscopic examination he described the characteristics of lesions in many tissues, which seemed to indicate that each disease was a specific alteration in the ratio of the two principal components of blood protein, fibrin and albumin. This imbalance was chemically based, he believed, and could be determined by chemical analysis of exudates and secretions, which themselves resulted from the stasis of blood flow produced by neurovascular damage.
This haemato-humoral theory of disease, based on anatomical and microscopic observations, suggested a chemical and neuropathological origin for disease and was consistent with the prevailing medical epistemology. William Harvey, John Hunter, Francois Magendi and Gabriel Andral all emphasized observation and accurate descriptions of structural changes in the human body as the basis of theory construction [24]. By contrast, Rokitansky's most vituperative critic, Rudolf Virchow, who was regarded as the father of modern scientific medicine, eschewed humoral, neural and vascular theories, and study of tissue lesions in favour of the individual cell as the structural and functional unit of disease.
While criticising Rokitansky, Virchow ignored the contemporary work in physiology of Claude Bernard. Although Bernard accepted Virchow's cellular theory of disease, his concept of the ‘milieu interieur'in no way necessitated such acceptance. The ‘milieu interieur’ describes the conditions necessary for sustaining life and that are regulated by activities such as chemical exchange, temperature control and the balance in the production of internal secretions. Neither anatomical pathology nor chemical histology could yield information about the functional relationship between tissues that Bernard claimed was the primary task of physiology. These relationships were necessary for life, and the role of experimental physiology was to alter them systematically by changing the conditions under which the relationships occur; this itself being Bernard's notion of disease: ‘physiology under altered environmental conditions’[25].
The point that the work of these three scientists illustrates is that each claimed to have made observations on which his theories were based, to have used observation to construct a classification of disease, and to have mapped out a programme for further research. The differences between the scientists lie in the ‘focus'of their observations, which depends on their implicit theories and on their views of the relationship between health and disease, the relationship between clinical and the experimental roles, and on different techniques of observation, instrumentation and experimental design.
When medicine introduces concepts of health and disease, new levels of complexity and abstraction emerge beyond those of the biological sciences. These higher-level concepts of clinical medicine are not givens, but reflect the reciprocal influences between scientific interpretation and a socially constructed reality; as a result they cause difficulty for empiricism [2], [26].
We cannot make sense of ‘facts'free from the influence of previous learning and a conceptual framework that selects and organizes the observations at different conceptual or epistemic levels. This need not be an albatross for a mature appreciation of and progress in clinical science. Developments in neuroscience illustrate this point.
Neuroscience
Texts of neuroanatomy and neurology traditionally describe centres of the brain that regulate its functions. Knowledge about the centres was inferred from observational evidence induced by injury, disease, experimental stimulation or surgical extirpation of parts of the brain. The resultant dysfunctions were classified as aphasias, apraxias, agnosias and alexias. The bewildering array of syndromes attests to the limitations of simple observation attempted without an underlying concept of brain function. Similar considerations bedevil the interpretation of contemporary neuroimaging studies.
The notion of simple localization derived from the pioneering work of Broca and Wernicke has been challenged and progressively modified since the late 19th century, beginning with Hughlings Jackson's hierarchical model of brain organization, which Freud combined with the idea that all mental representation involves associations between functionally different parts of the brain.
In a model informed by Husserl's phenomenology and classical Freudian psychoanalysis, Merleau-Ponty's phenomenological psychology emphasized the correspondence of psychoneural processes with the phenomenological field of perception and language [27]. He described how apparent deficits in neurological function could be viewed as adaptational responses of the damaged mind–body unity to its environment. Revision of the traditional localization model was also proposed by Geschwind in the light of observations of disconnection syndromes [28], whereas Luria's approach to e xperimental neuropsychology using principles of Gestalt psychology led him to construct a holistic model of the brain [29].
In yet another perspective, namely a synthesis of philosophical, psychoanalytic and Piagetian ideas about the nature of self-consciousness and acquisition of knowledge, Hundert [30] has proposed a model of brain function consistent with Edelman's theory of neuronal selection.
From studies of the putative subjective and intersubjective experiences of babies and their parents, Stern [31] has proposed a model of the mind based on neural networks and parallel distributed processes to explain the human capacity to integrate simultaneously occurring mental events, including motivational shifts, visual images, affective experiences, arousal states, sensations and memories. These may be linked in a narrative-like pattern of coherence that conveys meaning.
Siegel's ‘interpersonal neurobiology’ proposes that brain functions give rise to mind through several levels of integration from the neurological (right and left hemisphere integration) to the interpersonal (reciprocal emotional communication between a young child and its caregivers), each of which has its own methods of study [32]. Such integration underpins crucial developmental processes, including the consolidation of attachment, organization of emotional and self-defining experiences and the construction of meaning. These processes have been of fundamental interest to psychoanalysis for decades and are deemed by some to be pivotal in treating mental illness [33].
Psychosocial science
In an account of the epistemological status of generational cycles and their ‘developmental logic’, and how these relate to the Oedipus myth in different historical and cultural contexts, Erik Erikson described three elements that constitute our sense of reality: (i) the world of facts jointly recognized and verified during a particular period of cognitive development and in a particular technological period (factuality); (ii) the way human beings share and enthuse one another through exchanging these facts (actuality); and (iii) how facts are selected because they appear to ‘really exist'and are linked in a compelling context (contextuality) [34]. Erikson linked the latter to Einstein's aforementioned aesthetic considerations. He argued that this allows metaphysical, evolutionary and historical factors to be studied for their impact on individual development even though they cannot be directly observed.
Psychopathology
In defence of psychopathology as an independent basic science for psychiatry, Hoff draws on philosophy and experimental psychology to argue that neither ‘theory-free'empiricism nor a ‘mindless'passi ve reception of facts about the world is an adequate framework for any science, least of all for a science of psychopathology in which the human mind studies its own disturbances [35]. Instead, concepts of mental illness are created in psychiatry at several levels of data collection and interpretation.
Synthesis
There is no clear-cut distinction between liberalism and authoritarianism in modern science.
—Max Horkheimer Eclipse of Reason
All the above models are based on observation, but each is an interpretation of evidence derived from different assumptions about the subject, particular contexts of study and various modes of ‘observation’.
The fear of conceptual anarchy aroused by this pluralism dissolves once the inadequacies of the basic tenets of empiricism are appreciated and a more sophisticated formulation of medicine is proposed, such as the biopsychosocial model [36], which enables psychiatry to study clinically relevant phenomena at various levels of conceptual integration from the synaptic [37] to the social order [38], [39].
Thus research in attachment theory has ‘observ ed'that the unresolved/disorganized (u/d) category in the Adult Attachment Interview, which codes the lapses and incoherences in adults'narrati ves of their childhood losses or traumas correlates closely with children of such adults classified as belonging to the disorganized/disoriented category (group D) in the Ainsworth Infant Strange Situation [40]. Explanations for and implications of this correlation have been proposed by psychoanalysis, cognitive science, family systems theory, neurobiology and sociobiology. Both the attempt to integrate these explanations and to preserve their disciplinary autonomy greatly enrich psychiatry.
Teachers of psychiatry and planners of psychiatric services have failed to respect this epistemic pluralism that characterizes psychiatric knowledge, and to appreciate the clinical and ethical limitations of concepts derived from simple empiricism. Two recent examples of this failure are the deeply flawed concept of ‘serious mental illness’, which guided the first stage of the Australian government's National Mental Health Strategy [41] and its epistemic successor, the distinction between ‘high prevalence'and ‘low prevalence'mental disorders. Such failures may lead to claims of empirically derived, ‘evidence-based'treatments, the conceptual and clinical inadequacies of which remain unchallenged for years (42,43).
Conclusion
To the extent that the foundational subjects of psychiatry are scientific, empiricism is a conceptually naïve and historically inaccurate foundation. Marshalling and comparing a sufficient number of observable or measurable facts as evidence rarely, if ever, leads to the one correct conclusion. Intellectual flexibility, tolerance of ambiguous and discordant information obtained by different methods from differing viewpoints and at different conceptual levels, the judicious yet knowingly fallible, theory-derived construction, selection and interpretation of observations and empathically derived experiences typify the scientific method and is congruent with a form of clinical practice that is scientific, therapeutic and ethical.
Such understanding and practice are threatened by current trends in psychiatric training and the organization of services of which ‘evidence-based'approaches are predicated on a misplaced confidence in the power of empiricist principles to determine what constitutes reliable knowledge.
Footnotes
Acknowledgements
I thank Professor Tim van Gelder of the Department of Philosophy, University of Melbourne, for his comments on an earlier version of this paper.