On van Fraassen’s Constructive Empiricism


Constructive empiricism is a theory proposed by Bas van Fraassen (born 1941) in the book The Scientific Image, published in 1980. It is based on the idea that to hold a theory it is enough to agree that the theory is empirically adequate and it is not necessary believe that it describes the world as it actually is. To put it in other words, constructive empiricist thinks that empirical adequacy is the real quest for the science, not explanation of the world (Ladyman, p. 186). Or, as van Fraassen himself stated it, “the true demand of science is not for explanation as such, but for imaginative pictures which have a hope of suggesting new statements of observable regularities and of correcting old ones” (van Fraassen, p. 34).

Why this approach is worth to consider?

Some possible arguments in favor of this approach are.

The emergence of modern science. The first argument is based on the history of early modern science. To do so we should ask ourselves, what was the goal of the modern science in the very beginning? In a foreword for Copernicus’ De Revolutionibus, written by Andreas Osiander, it is stated: “…it is the duty of an astronomer to compose the history of the celestial motions through careful and expert study. Then he must conceive and devise the causes of these motions or hypotheses about them. Since he cannot in any way attain to the true causes, he will adopt whatever suppositions enable the motions to be computed correctly from the principles of geometry for the future as well as for the past. (…) For these hypotheses need not be true nor even probable. On the contrary, if they provide a calculus consistent with the observations, that alone is enough” (Copernicus). Van Fraassen mentions this fact as well in the beginning of his book: “The philosophers engaged in developing the philosophical foundations of modern science had apparently escaped this dilemma. Without postulating such causal properties, forms, or ‘occult qualities’, they could still explain the regularities that are observed in nature” (van Fraassen, p. 1).

Early modern astronomers and physicists did not regarded their works as a description of the real world. Empirical science, therefore, from the very beginning emerged as a study of empirical facts with no metaphysical claims. Only later effectiveness of science led to creation of ideology of scientism, which claimed that the science tells us about the real world. These claims, nevertheless, were not necessary for the emergence of modern science and were not considered as its goal. If so, why should we change this approach now? Moreover, van Fraassen compares realism with religious fundamentalism when he writes that the idea of “literal construal” of theories emerged in theology (van Fraassen, p. 10).

Occam’s razor. If empirical adequacy is perfectly enough for doing science, why should we add some sort of ontological claims? It seems that, at least for the sake of simplicity of our theories, we should not regard hypothetical objects of scientific theories as real. Such an approach would comply with “Occam’s razor” principle, which is often referred by scientists themselves as a prominent feature of rational approach to reality.

Practical consequences. We should also consider practical consequences of both approaches. If, according to §1, science can work with no ontological claims, it seems that it would be better not to make those claims, because ontological commitment to hypothetical entities can make science much more conservative and can encourage scientists to stick to their theories instead of changing them in accordance with new empirical evidence. In fact, at this point science can easily transform in a kind of religion with beliefs that are extremely important for its followers.

PMI. Finally, pessimistic meta-induction provides very sound support for any type of anti-realism, including van Fraassen’s constructive empiricism.

Based on these and other arguments, I am inclined to think that the most important proposition of constructive empiricism seems fairly rational and there are good reasons to agree that empirical adequacy, not revelation of any kind of metaphysical truth, is what is important for scientific theories.

Does distinction between observables and unobservables make sense?

Situation becomes worse, on the other hand, when we begin to examine particular aspects of van Fraassen’s approach. The idea of epistemic significance of observability as a criterion of ontological status of particular entities is questionable.

It is too anthropocentric approach. The idea of observability in this sense makes us to use a random and subjective criterion (whether something is available for our senses or not) to make ontological statements. At this point, I should agree with criticism provided by G. Maxwell who mentioned that “drawing of the observational theoretical line at any given point is an accident” (Ladyman, p. 187).

Unobservable observables. Even if we agree about the importance of observability, there is a huge and important class of objects that can be observed in principle, but still cannot actually be observed in real life. Consider dinosaurs as a good example of this kind of objects; it is quite confusing if these objects observable or not. Technically, we can observe them if and only if we invent a time machine or would be able to clone a dinosaur somehow, and the very possibility of these technologies is, in its turn, in question. This situation seems to be quite different from, say, an exoplanet, which is obviously observable in principle. Therefore, we cannot classify these objects as observable or unobservable.

Nevertheless, we can provide several arguments in favor of importance of observability. The most important argument in favor of constructive empiricism, in my opinion, is practical argument that the level of observables is the level where we usually operates in our everyday life. Because of this reason, any change on other level of reality is important for us if and only if it provides us with new possibilities to influence the level of observables. A humoral theory in medicine, for instance, was wrong, but it was so empirically successful when it comes to description of temperament types that we still use this classification of temperaments. That supports the idea that empiric adequacy is what we really look for in a scientific theory, while explanations are not so important. Nevertheless, this argument is far from van Fraassen’s own ideas, and when we return to constructive empiricism, we will easily see that it has another difficulties.

Is agnosticism about unobservables enough?

Although above mentioned arguments are far from van Fraassen’s own approach, they still provide good reasons to take the problem of observability seriously. Let us pretend to agree about the importance of observability. Nevertheless, there are at least two other important problems that arise from this idea.

“Atheism” about unobserables. Let us pretend that we agree with above mentioned arguments. In this situation why should not we choose to be rather “atheists” about unobservables than agnostics? Terms “agnosticism” and “atheism” are taken from Ladyman (p. 231). This term “atheism” is used to describe Larry Laudan’s approach based on implementation of IBE principle. As Ladyman put it, according to this view “…whatever the merits of IBE and whether or not constructive empiricism is ultimately defensible, scientific realism cannot be the best explanation of the success of science because it is not even empirically adequate” (Ibid.). This seems to be a very persuasive argument nor only against scientific realism, but also against constructive empiricism.

General agnosticism. Another tenet of constructive empiricism that can be put in question is the following: if we are agnostics about unobservables, why should not we also be agnostics about observable objects? Based on §5, it seems reasonable to expand our agnosticism and to become agnostics in a wider, say, Kantian manner instead of limit our agnosticism only by unobservable objects. Yet van Fraassen keeps strictly to draw a line between observables and unobservables which seems to be hard to defend.

Conclusions regarding constructive empiricism

Nevertheless, van Fraassen’s ideas should at least be considered very seriously. Van Fraassen’s argumentation may sometimes be weak, and his position is from time to time not consistent. However, we can elaborate other arguments in favor of the basic principle of van Fraassen’s anti-realism, according to which “science aims to give us theories which are empirically adequate; and acceptance of a theory involves as belief only that it is empirically adequate” (van Fraassen, p. 12).


  1. Copernicus, N. (1543) De Revoltionibus. Transl. and comment. by Edward Rosen.
  2. van Fraassen, B. (1980) The Scientific Image. E-book from Oxford Scholarship Online.
  3. Ladyman, J. (2002) Understanding Philosophy of Science, Routledge, London.
  4. Monton, B., Mohler, Ch. Constructive Empiricism. In: Stanford Encyclopedia of Philosophy.

Scientism as a Case of Pseudoscience

When we talk about a pseudoscience, one usually think about religion-related examples like creation science or astrology as an example of pseudoscience. However, pseudoscience is often generated by scientists themselves when they try to make universal implications of their particular scientific knowledge. As an example, let us take scientism.

2016-04-25 · Read ›