Psychoanalysis and science

Introduction

The debate over the scientific status of psychoanalysis, which began with its birth at the end of the nineteenth century, has now extended well into the twenty-first. Bion, taking an expansive view of what constituted science, wrote that all investigation is ultimately scientific. As we have seen, he designated the activity of acquiring knowledge by any means whatsoever, as well as the desire or impulse behind this activity, by the letter K for knowledge. In this broad view, there are many different kinds of activity that satisfy the notion of K, from the most mathematically formal and elegant theories of physics, to what would generally be considered purely artistic productions. However much they may differ, these activities all convey some truth about the world.

When we hear the word “science”, however, our vision tends to shrink until we are left only with an image of a chemist in a white smock working in a laboratory full of glassware and complex instruments, or a physicist working in a cave in Switzerland filled with multi-billion dollar instruments for breaking atoms and analyzing the debris. This type of science—so-called hard science— produces results and theoretical predictions that are detailed, precise and highly reliable. The elegance and power of these results often blind us to the fact that they constitute only a small part of scientific activity. Even in the “hard” sciences themselves, the proportion of scientific problems that can be solved by such rigorous methods is quite small. Chemists and physicists not only do not have solutions to most problems in the chemical and physical world, they do not even know how to formulate these problems. The range of our ignorance is far vaster than the range of our knowledge.¹

Replicability: theories in experimental science and in psychoanalysis

Scientific theories that do yield precise and elegant predictions do so only because they have been repeatedly verified by rigorously controlled experimentation that produces results amenable to statistical analysis. Statistical analysis can produce results that are highly reliable and reproducible, but the reliability of these results depends on having a large number of examples to feed into the analysis. Such experimental studies are therefore confined to phenomena that are replicable. “Replicable” means that a great number of identical examples of the phenomenon in question may be studied simultaneously. The key words here are “great number”, and “identical”, which means that the phenomena one may study by such rigorous methods are necessarily simple.

Once established, theories in the hard sciences have predictive power that allows them to replace direct experience in future specific instances of the phenomena to which the theories apply. For example, Bernoulli’s Principle, which began as an hypothesis about the relationship between the movement of a fluid and the pressure it exerts on adjacent surfaces, has been so fully verified by controlled experiment that engineers now use it to predict very precisely how much lift a certain wing design will produce. Aircraft designers need not, therefore, construct a series of wings and test them by trial and error in order to know how to build a wing with the desired amount of lift.

In sharp contrast to disciplines that can produce these general, rigorously testable theories are disciplines that study phenomena that are too complex for such formalization, an example—perhaps an extreme example—of which is psychoanalysis. The requirement of hard science for a high degree of replicability excludes psychoanalytic studies of the human mind. We cannot replicate either ourselves or our patients.

Because psychoanalytic hypotheses cannot be confirmed by rigorous experiment, they cannot be turned into general theories that can replace direct experience in specific cases the way that the theories of the experimental sciences, such as Bernoulli’s Principle, can. On the contrary, to the degree that the analyst tries to emulate the aeronautical engineer by attempting to use psychoanalytic theory in place of direct trial and error experience with his patient, he falls short of analyzing the patient.

Psychoanalysts may recognize in a patient’s experience an instance of some general theory or piece of knowledge he already has—“this is splitting” (or Oedipal conflict, or denial, or reaction formation, and so on). But the analyst who merely applies theories—who fails to move himself (or be moved by his experience of the analysis) beyond what was already encompassed by his theories before his encounter with the patient—is courting analytic sterility.² He cannot simply assume that the clinical problem at hand represents a specific instance of a general law, and apply the general law to the specific instance, since the loss of specific, idiosyncratic detail that appeal to a general theory would entail would have a disastrous effect on the appropriateness (life-likeness) of the resulting interpretation. He can arrive at his interpretation—which is a kind of ad hoc theory about the patient—only by patiently absorbing as much as he can of the detail that is unique and specific to his clinical experience with one patient.

For this reason, the analyst must maintain a state of highly polished ignorance about what the patient presents him with, until his experience of the patient impresses something on him. An aeronautical engineer who tried to design an aeroplane in this way would be a very bad engineer, having to re-invent the aeroplane the way the Wright brothers did, by trial and error, each time. An analyst who did not proceed in this way would be a very bad analyst. The psychoanalyst, for practical purposes, must re-invent the wing each time she makes an interpretation: she must “forget” her psychoanalytic theory until the patient reminds her of it.

For problems in the study of the human mind, another kind of knowledge, less precise, and less certain, is appropriate. The “theories” that express this knowledge may hardly deserve the term “theory” at all. They are working hypotheses that are informal, highly specific and subject to immediate correction.

In light of these contrasts, one might suppose that the work of the experimental scientists and the work of the psychoanalyst could hardly be more different. But one would be wrong. Examples of this informal and temporary kind of theory abound in the experimental sciences, for example the hunches or guesses a scientist makes when exploring a new area in which he is still trying to get his bearings. (An outstanding example of this is Watson and Crick’s inspired guess about the molecular structure of DNA, based on X-ray data that was in itself very far from nailing down a specific structure.) This type of theory is always a work in progress. Highly informal, on-the-fly, and ad hoc, it would not be far from the mark to call it guesswork. It is far more common in the day-to-day activities of scientists, at least when they are working creatively, than is the production of rigorous, formal theories. This fluid guessing is found not only in the work of creative scientists and psychoanalysts, but also in the work of skilled craftsmen, where it takes the form of a combination of competence and curiosity and is addressed to the solution of a unique and specific problem. This know-how or savior-faire may be less prestigious than the rigor and precision of formal science, but the skilled craftsman has his own form of knowledge that is part of Bion’s “K”.

Formal theories in the experimental sciences are established by the construction of controlled experiments, while the informal theories of both physical science and psychoanalysis are established by the digestion of uncontrolled experience.