topeesp.gif (5672 bytes)

[ Scientific Activities - Actividades Científicas ]

Biomedical Science: an Epistemological point of view

Dr. Ricardo H. Pichel

The hypothetical-deductive method and "simple systems"
Complex systems and empirical-inductive methods
Recommended Reading

Whatever the science we are dealing with, there is a common starting point for all of them consisting of the set of "observables", which compose their empirical base. As concrete examples, the pendulum, the movement of the planets and colisions between bodies are "observables" found in the empirical basis of physics just as the protozoa, the invertebrates, the mammals and the birds are part of the observables found in the empirical basis of biology.

They are designated as such precisely because directly or indirectly they are available to "observation" by those who are interested in them.

We must first make clear what we expect from scientific knowledge seen from the epistemological point of view. The first is that it should be capable of explaining, that is to say, that it permits us to establish some kind of relationship between the events such that if one of them takes place the other must necessarily follow upon it. In second place, such an explanation should be as embracing as possible, in other words, the explanation should unify events which apparently are unconnected. Finally, it should have a predictive capability, that is to say, by a correct deduction it is possible to predict events or phenomena capable of being challenged experimentally in order to verify the validity of that prediction.

The first thing that we expect from a science is that it shows certain "observable regularities"; that is to say, that it allows us to identify within its empirical basis which things happen with a certain "regularity" and which things do not.

The hypothetical-deductive method and "simple systems"

At this stage, it is worth mentioning that there exist at least two classes of "observables": those which by their "simple" nature permit the science which concerns them to comply satisfactorily with the epistemological expectations we have defined above, that is to say, those which explain, unify and predict everything about these "observables" and, on the other hand, those which due to their enormous degree of complexity, do not always permit those sciences which concern them to achieve these three objectives.

Additionally, it is convenient to mention in advance that the hypothetical-deductive method is the one which permits us to comply with the epistemological objectives which we have imposed (explain, unify and predict), with the greatest degree of logical guarantee, and which, like none other, is bestowed with that degree of epistemological robustness.

It is evident at first glance that "simple systems" are susceptible to approach by the hypothetical-deductive method and that this approach appears to be the most appropriate for "simple systems".

As an exercise, let us begin to imagine what would have happened to physics had Isaac Newton not existed.

In pre-newtonian physics, the pendulum, the movement of the planets and colisions between bodies formed three types of "observable regularities" absolutely independent of each other. The laws of the pendulum, had they existed, would have been the result of careful measurements which would have revealed the "regularity" relating length to period: however, the inevitable friction of the different pendula would have conspired against the precision of these "regularities" which would therefore have been assigned a number indicating the probability that such a "regularity" was really present. Obviously, due to the simple nature of the pendulum, such probability would have been very high.

On the other hand, the "regularity" associated with colisions between bodies would have been limited to a statement of each case in particular, and considering that the concept of mass did not exist, any attempt to relate weights, velocities and directions would have suffered such an enormous limitation, that the possibility of predicting the course of a colision would have been reduced to a mere "singularity".

Referring to the movement of the planets, Kepler had stated his laws and was to a certain extent the individual who had best described such "regularity".

If this situation had continued, physics would have failed to comply with the three cardinal aims of science: a) explanation, b) unification, and c) prediction.

If this situation had continued, physics would be nothing more than a simple catalog of "observable regularities" accompanied by numbers indicating the probabilities of such "regularities". This, and only this, is the achievement of the empirical-inductive method.

What did Newton achieve?

Newton discovered the world of the "unobservable" and "created" three principles which he "decreed" to be true, and from which not only explained and unified the pendulum with the movement of the planets and colisions between bodies demonstrating that each of these obeyed the same laws, but also "predicted" the future behavior that a given system would show, knowing its initial conditions.

Without knowing it, Newton had "created" the hypothetical-deductive method, which as we have said, is the most appropriate one for the study of "simple systems".

In effect, the power of this method lies in the strong logical support that it incorporates.

Let us imagine for a moment that proposition A belongs to the domain of the "unobservables" and, as such, I am unable to prove if the proposition is "true" or "false".

It so happens that "correct" reasoning can never originate from truth and conclude in falseness. The "preservation" of truth is a property of "correct" reasoning which guarantees the hypothetical-deductive method.

Let us return to our proposition A. We said that, due to it being an "unobservable", we cannot establish if it is true or false, but let us suppose that through application of "correct" reasoning and starting at A, we reach proposition B, which fortunately belongs to the domain of the "observables", and as such, we can apply to it methods which are capable of establishing whether it is true or false. Suppose that we apply these methods and conclude that proposition B is true; are we now able to conclude something about the truth or falseness of A?. Absolutely not, because even though we have applied correct reasoning and, as such, we can be sure that we have not gone from truth to falseness, the case may well be that we have gone from falseness to truth or that we have gone from truth to truth, which leads us to conclude that although B may be demonstrably true, we can never use this property of B to demonstrate the nature of A. This is the crux of the hypothetical-deductive method. What would happen if we now "decree" that A is true and apply correct reasoning to predict the nature of B? If we find that B is true, we still cannot say anything about A, as before. We can only conclude that what we have decreed about A has not been proved wrong. However, if we find that B is False, then we can conclude definitively that A is also false. This is the power of the method.

As we have seen, the hypothetical-deductive method is the one that best explains, unifies, and predicts in the case of simple systems. In physics this method has gained great prestige because it is physics which has most commonly studied simple systems, thus converting itself into a paradigm for this method, and it is physics which has gained most ground for epistemology. The Great Laws of Nature are the starting points for investigation in physics and the robustness of theoretical physics reveals that a good part of what physicists do is to use these laws, or to invent others capable of explaining phenomena beyond the reach of these.

Complex systems and empirical-inductive methods.

Let us start by stating that there are varying degrees of complexity and that biological systems are amongst the most complex of all.

Complex systems are curious in the sense that not only is it difficult to define a starting point from which to deduce their properties and so subject them to experimental verification, but also that they are not amenable to explanation using the Great Laws of Nature, which were obtained from, and are capable of explaining, a multiplicity of simple systems.

For one reason or another, complex systems, and in particular, biological systems, have been approached using methods which are fundamentally morphological and descriptive, and in the best of cases, by inductive-empirical methods whose epistemological value does not rise above that of "observable regularities" associated with not-too significant statistics. The description of pre-newtonian physics made at the beginning of this article is perfectly applicable to our present state of knowledge about complex systems.

In our knowledge of complex systems we can only say that there are "findings", findings which are supported only by a statistical evaluation which guarantees that the finding is indeed a "regularity", bereft of any explanation capable of providing unification, or predictability other than that which arises from a statistical indication that the probability of a given event occurring is greater when a certain set of conditions is present.

I believe that the problem we have in the study of complex systems is related, on one hand, to their very nature (which is a legitimate concern), and on the other, to the difficulty of overcoming an experimentalist tradition, which is intuitive and which stubbornly clings to "reality", resisting any attempt to simplify the system in order to develop a model based only on the most outstanding variables. This epistemological posture reassures our belief that what we are describing is indeed an observable regularity, but does not permit any attempt to include in a description of the system any factor which is not the subject of rigorous scientific proof, even though such factor might explain the observed regularity and might indeed permit predictions about the system. This type of approach is considered to transgress scientific rigor and is rejected by most peer reviewers.

Notable exceptions to this posture which have become part of history are the ideas of Gregor Mendel, who proposed the existence of genes to explain and predict heredity, and Charles Darwin, who developed the theory of evolution to explain the appearance and extinction of different animal species. Nonetheless, the general tendency in the study of complex systems is to cling to inductivist intuition, which blindly confides in the observable and is blind to ideas which arise from the world of the unobservable.

If physics had been left in the hands of biomedical scientists, we would still today rely on candles for illumination because they could not see the electron and would therefore refuse to formulate any type of law concerning such an unrealistic substrate. Fortunately, Ohm, Faraday and Maxwell were not biomedical scientists.

In my personal experience, the first difficulty which I have consistently met happens when we try to develop a model in which we include only those variables thought to be more important than others found in the system. This process of "excluding" variables is seen as cheating on "reality" instead of being seen as an attempt to simplify the process in order to find out if the chosen variables are capable of explaining the observed events and predicting future ones. If this can be done to a good degree of approximation, then those variables which were excluded deserved to be so, and if not, there is no doubt that relevant variables were excluded or inappropriate ones included. It is not true to say that simplification violates the process under study; on the contrary, not to do so denies freedom of action.

Even in the case of simple systems, exclusion of variables has been important. Friction was present in Newton’s experiments, but if he had not imagined a world without friction, he would never had been able to formulate the principle of inertia.

If we continue to obstinately resist exclusion of variables, thus unnecessarily complicating our investigations, we will remain at the level of merely describing a phenomenon (macroscopically, microscopically, or even at the molecular level) without a degree of understanding which allows us to explain, unify and predict beyond this.

Over the last few years, complex systems have attracted the attention of many physicists and mathematicians and, fortunately, they have found that biological systems respect the same formalisms used in the analysis of complex systems of other natural phenomena.

I have great confidence in this new approach, but I fear that while it matures we will see more examples of what is already happening in some allied disciplines, such as psychology, particularly the Lacanian school, which interprets in its own way concepts which the physics of complex systems is still far from proving.

Whenever a new wide-ranging field is explored, such as the physics of complex systems, there exists a legion of dilettantes characterized by their precarious understanding, who wish to add grist to their respective ideological mills. I will never forget an article written by Vargas Llosa which appeared in the Cultural Supplement of La Nacion under the title "Welcome Chaos". On the basis of some conversations that he had held in some part of the world with a French physicist who was studying chaos theory, Don Mario understood what he could of chaos, but enough for him to conclude that the scientific basis of capitalism could be explained by it.


Let us start by making clear some points that have been confusing right from the beginnings of medicine.

Medical doctors are concerned with the most complex system which exists in nature: man himself. As a complex system, Man is in a constant state of cognoscibility. That is to say, there is still much to be known about him and so it is not possible to think of medicine divorced from scientific investigation directed both to man and to complex systems in general.

We may be more or less in disagreement with the epistemological viewpoint of the biomedical sciences, but we cannot disagree with the use of epistemology in this context. If we take as an example the range of products which have been developed by man thanks to the application of the General Laws and the degree of technical sophistication achieved in them, and we compare them with complex systems in which man has had no intervention, products of natural evolution or of God, we can easily recognise the enormous difference in attitude we show towards them. Is it worth investigating a TV set? Obviously not; it is the product of a human mind and everything is known about it. The only thing left is to improve upon it, make it sharper, make a three-dimensional image, using our knowledge extracted from simple systems which have formed the building blocks of the general laws of nature.

Is it worth investigating Man or complex systems? Obviously it is, because man did not intervene in their formation and as products of God or natural evolution they represent the truly unknown and thus a primary challenge for science.

From the medical point of view, there exists a motivation which does not involve science as such, being based on the need to "modify" the human being when he is suffering an illness. This "modification" does indeed require knowledge, and the doctor knows that he must inevitably search for it in science, but it should be firmly understood that the motivation for the doctor is modification and not understanding for its own sake.

For the doctor there exists a criterion of "pertinence" of knowledge, which the scientist does not have. The scientist is more interested in understanding whereas the doctor is more interested in using the knowledge so that he can modify the situation of the patient.

For the scientist, what matters most is the "quality" of a given piece of knowledge whereas the doctor is only interested in quality knowledge which can be applied in his profession.

Having made these points clear, our criticism of the biomedical sciences does not therefore include the doctor but does include the investigators who work in the biomedical field. The doctor relies upon the discoveries of the scientists, and he uses only those which help him to "modify" his patient.

The doctor, however, knows more than anyone else that he is working on a complex system and for this reason when he applies research-based knowledge he is very careful in observing the response of this unexplored complexity, and without necessarily knowing why or how he decides to carry out a "trial" of any new resource, comparing its results with a control group.

The clinical trial is not a method of scientific investigation, it is more of a precaution which the doctor takes to "see" if the "proposed modification" generates unexpected effects in a system whose unapproachable complexity does not allow him to make predictions using the classical methods of investigation.

Recommended Reading:

Carnap, R., Fundamentación Lógica de la Física, Madrid, Hyspamérica, 1985.
Nagel, E. , La Estructura de la Ciencia, Buenos Aires, Paidós, 1968.
Popper, K., La Lógica de la Investigación Científica, Madrid, Tecnos, 1973.
Klimovsky, G., Las Desventuras del Conocimiento Científico, Buenos Aires, A-Z Editora, 1994.